Israeli Online Attack Service ‘vDOS’ Earned $600,000 in Two Years

SEP 16

Israeli Online Attack Service ‘vDOS’ Earned $600,000 in Two Years

vDOS  a “booter” service that has earned in excess of $600,000 over the past two years helping customers coordinate more than 150,000 so-called distributed denial-of-service (DDoS) attacks designed to knock Web sites offline — has been massively hacked, spilling secrets about tens of thousands of paying customers and their targets.

The vDOS database, obtained by at the end of July 2016, points to two young men in Israel as the principal owners and masterminds of the attack service, with support services coming from several young hackers in the United States.

The vDos home page.

The vDos home page.

To say that vDOS has been responsible for a majority of the DDoS attacks clogging up the Internet over the past few years would be an understatement. The various subscription packages to the service are sold based in part on how many seconds the denial-of-service attack will last. And in just four months between April and July 2016, vDOS was responsible for launching more than 277 million seconds of attack time, or approximately 8.81 years worth of attack traffic.

Let the enormity of that number sink in for a moment: That’s nearly nine of what I call “DDoS years” crammed into just four months. That kind of time compression is possible because vDOS handles hundreds — if not thousands — of concurrent attacks on any given day.

Although I can’t prove it yet, it seems likely that vDOS is responsible for several decades worth of DDoS years. That’s because the data leaked in the hack of vDOS suggest that the proprietors erased all digital records of attacks that customers launched between Sept. 2012 (when the service first came online) and the end of March 2016.


The hack of vDOS came about after a source was investigating a vulnerability he discovered on a similar attack-for-hire service called PoodleStresser. The vulnerability allowed my source to download the configuration data for PoodleStresser’s attack servers, which pointed back to api.vdos-s[dot]com. PoodleStresser, as well as a large number of other booter services, appears to rely exclusively on firepower generated by vDOS.

From there, the source was able to exploit a more serious security hole in vDOS that allowed him to dump all of the service’s databases and configuration files, and to discover the true Internet address of four rented servers in Bulgaria (at that are apparently being used to launch the attacks sold by vDOS. The DDoS-for-hire service is hidden behind DDoS protection firm Cloudflare, but its actual Internet address is

vDOS had a reputation on cybercrime forums for prompt and helpful customer service, and the leaked vDOS databases offer a fascinating glimpse into the logistical challenges associated with running a criminal attack service online that supports tens of thousands of paying customers — a significant portion of whom are all trying to use the service simultaneously.

Multiple vDOS tech support tickets were filed by customers who complained that they were unable to order attacks on Web sites in Israel. Responses from the tech support staff show that the proprietors of vDOS are indeed living in Israel and in fact set the service up so that it was unable to attack any Web sites in that country — presumably so as to not attract unwanted attention to their service from Israeli authorities. Here are a few of those responses:

(‘4130′,’Hello `d0rk`,\r\nAll Israeli IP ranges have been blacklisted due to security reasons.\r\n\r\nBest regards,\r\nP1st.’,’03-01-2015 08:39),

(‘15462′,’Hello `g4ng`,\r\nMh, neither. I\’m actually from Israel, and decided to blacklist all of them. It\’s my home country, and don\’t want something to happen to them :)\r\n\r\nBest regards,\r\nDrop.’,’11-03-2015 15:35),

(‘15462′,’Hello `roibm123`,\r\nBecause I have an Israeli IP that is dynamic.. can\’t risk getting hit/updating the blacklist 24/7.\r\n\r\nBest regards,\r\nLandon.’,’06-04-2015 23:04),

(‘4202′,’Hello `zavi156`,\r\nThose IPs are in israel, and we have all of Israel on our blacklist. Sorry for any inconvinience.\r\n\r\nBest regards,\r\nJeremy.’,’20-05-2015 10:14),

(‘4202′,’Hello `zavi156`,\r\nBecause the owner is in Israel, and he doesn\’t want his entire region being hit offline.\r\n\r\nBest regards,\r\nJeremy.’,’20-05-2015 11:12),

(‘9057′,’There is a option to buy with Paypal? I will pay more than $2.5 worth.\r\nThis is not the first time I am buying booter from you.\r\nIf no, Could you please ask AplleJack? I know him from Israel.\r\nThanks.’,’21-05-2015 12:51),

(‘4120′,’Hello `takedown`,\r\nEvery single IP that\’s hosted in israel is blacklisted for safety reason. \r\n\r\nBest regards,\r\nAppleJ4ck.’,’02-09-2015 08:57),


As we can see from the above responses from vDOS’s tech support, the owners and operators of vDOS are young Israeli hackers who go by the names P1st a.k.a. P1st0, and AppleJ4ck. The two men market their service mainly on the site hackforums[dot]net, selling monthly subscriptions using multiple pricing tiers ranging from $20 to $200 per month. AppleJ4ck hides behind the same nickname on Hackforums, while P1st goes by the alias “M30w” on the forum.

Some of P1st/M30W's posts on Hackforums regarding his service vDOS.

Some of P1st/M30W’s posts on Hackforums regarding his service vDOS.

vDOS appears to be the longest-running booter service advertised on Hackforums, and it is by far and away the most profitable such business. Records leaked from vDOS indicate that since July 2014, tens of thousands of paying customers spent a total of more than $618,000 at the service using Bitcoin and PayPal.

Incredibly, for brief periods the site even accepted credit cards in exchange for online attacks, although it’s unclear how much the site might have made in credit card payments because the information is not in the leaked databases.

The Web server hosting vDOS also houses several other sites, including huri[dot]biz,ustress[dot]io, and vstress[dot]net. Virtually all of the administrators at vDOS have an email account that ends in v-email[dot]org, a domain that also is registered to an Itay Huri with a phone number that traces back to Israel.

The proprietors of vDOS set their service up so that anytime a customer asked for technical assistance the site would blast a text message to six different mobile numbers tied to administrators of the service, using an SMS service called Two of those mobile numbers go to phones in Israel. One of them is the same number listed for Itay Huri in the Web site registration records for v-email[dot]org; the other belongs to an Israeli citizen named Yarden Bidani. Neither individual responded to requests for comment.

The leaked database and files indicate that vDOS uses Mailgun for email management, and the secret keys needed to manage that Mailgun service were among the files stolen by my source. The data shows that vDOS support emails go to itay@huri[dot]biz, and


The $618,000 in earnings documented in the vDOS leaked logs is almost certainly a conservative income figure. That’s because the vDOS service actually dates back to Sept 2012, yet the payment records are not available for purchases prior to 2014. As a result, it’s likely that this service has made its proprietors more than $1 million.

vDOS does not currently accept PayPal payments. But for several years until recently it did, and records show the proprietors of the attack service worked assiduously to launder payments for the service through a round-robin chain of PayPal accounts.

They did this because at the time PayPal was working with a team of academic researchers to identify, seize and shutter PayPal accounts that were found to be accepting funds on behalf of booter services like vDOS. Anyone interested in reading more on their success in making life harder for these booter service owners should check out my August 2015 story, Stress-Testing the Booter Services, Financially.

People running dodgy online services that violate PayPal’s terms of service generally turn to several methods to mask the true location of their PayPal Instant Payment Notification systems. Here is an interesting analysis of how popular booter services are doing so using shell corporations, link shortening services and other tricks.

Turns out, AppleJ4ck and p1st routinely recruited other forum members on Hackforums to help them launder significant sums of PayPal payments for vDOS each week.

“The paypals that the money are sent from are not verified,” AppleJ4ck says in one recruitment thread. “Most of the payments will be 200$-300$ each and I’ll do around 2-3 payments per day.”

vDos co-owner AppleJ4ck recruiting Hackforums members to help launder PayPal payments for his booter service.

vDos co-owner AppleJ4ck recruiting Hackforums members to help launder PayPal payments for his booter service.

It is apparent from the leaked vDOS logs that in July 2016 the service’s owners implemented an additional security measure for Bitcoin payments, which they accept through Coinbase. The data shows that they now use an intermediary server ( to handle Coinbase traffic. When a Bitcoin payment is received, Coinbase notifies this intermediary server, not the actual vDOS servers in Bulgaria.

A server situated in the middle and hosted at a U.S.-based address from Digital Oceanthen updates the database in Bulgaria, perhaps because the vDOS proprietors believed payments from the USA would attract less interest from Coinbase than huge sums traversing through Bulgaria each day.


The extent to which the proprietors of vDOS went to launder profits from the service and to obfuscate their activities clearly indicate they knew that the majority of their users were using the service to knock others offline.

Defenders of booter and stresser services argue the services are legal because they can be used to help Web site owners stress-test their own sites and to build better defenses against such attacks. While it’s impossible to tell what percentage of vDOS users actually were using the service to stress-test their own sites, the leaked vDOS logs show that a huge percentage of the attack targets are online businesses.

In reality, the methods that vDOS uses to sustain its business are practically indistinguishable from those employed by organized cybercrime gangs, said Damon McCoy, an assistant professor of computer science at New York University.

In reality, the methods that vDOS uses to sustain its business are practically indistinguishable from those employed by organized cybercrime gangs, said Damon McCoy, an assistant professor of computer science at New York University.

“These guys are definitely taking a page out of the playbook of the Russian cybercriminals,” said McCoy, the researcher principally responsible for pushing vDOS and other booter services off of PayPal (see the aforementioned story Stress-Testing the Booter Services, Financially for more on this).

“A lot of the Russian botnet operators who routinely paid people to infect Windows computers with malware used to say they wouldn’t buy malware installs from Russia or CIS countries,” McCoy said. “The main reason was they didn’t want to make trouble in their local jurisdiction in the hopes that no one in their country would be a victim and have standing to bring a case against them.”

The service advertises attacks at up to 50 gigabits of data per second (Gbps). That’s roughly the equivalent of trying to cram two, high-definition Netflix movies down a target’s network pipe all at the same moment.

But Allison Nixon, director of security research at business risk intelligence firmFlashpoint, said her tests of vDOS’s service generated attacks that were quite a bit smaller than that — 14 Gbps and 6 Gbps. Nevertheless, she noted, even an attack that generates just 6 Gbps is well more than enough to cripple most sites which are not already protected by anti-DDoS services.

And herein lies the rub with services like vDOS: They put high-powered, point-and-click cyber weapons in the hands of people — mostly young men in their teens — who otherwise wouldn’t begin to know how to launch such attacks. Worse still, they force even the smallest of businesses to pay for DDoS protection services or else risk being taken offline by anyone with a grudge or agenda.

“The problem is that this kind of firepower is available to literally anyone willing to pay $30 a month,” Nixon said. “Basically what this means is that you must have DDoS protection to participate on the Internet. Otherwise, any angry young teenager is going to be able to take you offline in a heartbeat. It’s sad, but these attack services mean that DDoS protection has become the price of admission for running a Web site these days.”

Stay tuned for the next piece in this series on the hack of vDOS, which will examine some of the more interesting victims of this service.

From damage to discovery via virtual unwrapping: Reading the scroll from En-Gedi

From damage to discovery via virtual unwrapping: Reading the scroll from En-Gedi
Science Advances  21 Sep 2016:
Vol. 2, no. 9, e1601247
DOI: 10.1126/sciadv.1601247


Computer imaging techniques are commonly used to preserve and share readable manuscripts, but capturing writing locked away in ancient, deteriorated documents poses an entirely different challenge. This software pipeline—referred to as “virtual unwrapping”—allows textual artifacts to be read completely and noninvasively. The systematic digital analysis of the extremely fragile En-Gedi scroll (the oldest Pentateuchal scroll in Hebrew outside of the Dead Sea Scrolls) reveals the writing hidden on its untouchable, disintegrating sheets. Our approach for recovering substantial ink-based text from a damaged object results in readable columns at such high quality that serious critical textual analysis can occur. Hence, this work creates a new pathway for subsequent textual discoveries buried within the confines of damaged materials.


  • Micro-CT
  • virtual unwrapping
  • segmentation
  • digital restoration
  • visualization
  • digital flattening


In 1970, archeologists made a dramatic discovery at En-Gedi, the site of a large, ancient Jewish community dating from the late eighth century BCE until its destruction by fire circa 600 CE. Excavations uncovered the synagogue’s Holy Ark, inside of which were multiple charred lumps of what appeared to be animal skin (parchment) scroll fragments (12). The Israel Antiquities Authority (IAA) faithfully preserved the scroll fragments, although in the 40 years following the discovery, no one produced a means to overcome the irreversible damage they had suffered in situ. Each fragment’s main structure, completely burned and crushed, had turned into chunks of charcoal that continued to disintegrate every time they were touched. Without a viable restoration and conservation protocol, physical intervention was unthinkable. Like many badly damaged materials in archives around the world, the En-Gedi scroll was shelved, leaving its potentially valuable contents hidden and effectively locked away by its own damaged condition (Fig. 1).

Fig. 1The charred scroll from En-Gedi.

Image courtesy of the Leon Levy Dead Sea Scrolls Digital Library, IAA. Photo: S. Halevi.


The implementation and application of our computational framework allows the identification and scholarly textual analysis of the ink-based writing within such unopened, profoundly damaged objects. Our systematic approach essentially unlocks the En-Gedi scroll and, for the first time, enables a total visual exploration of its interior layers, leading directly to the discovery of its text. By virtually unwrapping the scroll, we have revealed it to be the earliest copy of a Pentateuchal book ever found in a Holy Ark. Furthermore, this work establishes a restoration pathway for damaged textual material by showing that text extraction is possible while avoiding the need for injurious physical handling. The restored En-Gedi scroll represents a significant leap forward in the field of manuscript recovery, conservation, and analysis.

Virtual unwrapping

Our generalized computational framework for virtual unwrapping applies to a wide range of damaged, text-based materials. Virtual unwrapping is the composite result of segmentation, flattening, and texturing: a sequence of transformations beginning with the voxels of a three-dimensional (3D) unstructured volumetric scan of a damaged manuscript and ending with a set of 2D images that reveal the writing embedded in the scan (36). The required transformations are initially unknown and must be solved by choosing a model and applying a series of constraints about the known and observable structure of the object. Figure 2 shows the final result for the scroll from En-Gedi. This resultant image, which we term the “master view,” is a visualization of the entire surface extracted from within the En-Gedi scroll.

Fig. 2Completed virtual unwrapping for the En-Gedi scroll.


The first stage, segmentation, is the identification of a geometric model of structures of interest within the scan volume. This process digitally recreates the “pages” that hold potential writing. We use a triangulated surface mesh for this model, which can readily support many operations that are algorithmically convenient: ray intersection, shape dynamics, texturing, and rendering. A surface mesh can vary in resolution as needed and forms a piecewise approximation of arbitrary surfaces on which there may be writing. The volumetric scan defines a world coordinate frame for the mesh model; thus, segmentation is the process of aligning a mesh with structures of interest within the volume.

The second stage, texturing, is the formation of intensity values on the geometric model based on its position within the scan volume. This is where we see letters and words for the first time on the recreated page. The triangulated surface mesh offers a direct approach to the texturing problem that is similar to solid texturing (78): Each point on the surface of the mesh is given an intensity value based on its location in the 3D volume. Many approaches exist for assigning intensities from the volume to the triangles of the segmented mesh, some of which help to overcome noise in the volumetric imaging and incorrect localization in segmentation.

The third stage, flattening, is necessary because the geometric model may be difficult to visualize as an image. Specifically, if text is being extracted, it will be challenging to read on a 3D surface shaped like the cylindrical wraps of scrolled material. This stage solves for a transformation that maps the geometric model (and associated intensities from the texturing step) to a plane, which is then directly viewable as a 2D image for the purpose of visualization.

In practice, this framework is best applied in a piecewise fashion to accurately localize a scroll’s highly irregular geometry. Also, the methodology required to map each of these steps from the original volume to flattened images involves a series of algorithmic decisions and approximations. Because textual identification is the primary goal of our virtual unwrapping framework, we tolerate mathematical and geometric error along the way to ensure that we extract the best possible images of text. Hence, the final merging and visualization step is significant not only for composing small sections into a single master view but also for checking the correctness and relative alignments of individual regions. Therefore, it is crucial to preserve the complete transformation pipeline that maps voxels in the scan volume to final pixels in the unwrapped master view so that any claim of extracted text can be independently verified.

The volumetric scan

The unwrapping process begins by acquiring a noninvasive digitization that gives some representation of the internal structure and contents of an object in situ (911). There are a number of choices for noninvasive, penetrative, and volumetric scanning, and our framework places no limits on the modality of the scan. As enhancements in volumetric scanning methodology [for example, phase-contrast microtomography (612)] occur, we can take advantage of the ensuing potential for improved images. Whatever the scanning method, it must be appropriate to the scale and to the material and physical properties of the object.

Because of the particularities of the En-Gedi scroll, we used x-ray–based micro–computed tomography (micro-CT). The En-Gedi scroll’s damage creates a scanning challenge: How does one determine the correct scan protocol before knowing how ink will appear or even if the sample contains ink at all? It is the scan and subsequent pipeline that reveal the writing. After several calibration scans, a protocol was selected that produced a visible range of intensity variation on the rolled material. The spatial resolution was adjusted with respect to the sample size to capture enough detail through the thickness of each material layer to reveal ink if present and detectable. The chemical composition of the ink within the En-Gedi scroll remains unknown because there are no exposed areas suitable for analysis. However, the ink response within the micro-CT scan is denser than other materials, implying that it likely contains metal, such as iron or lead.

Any analysis necessitates physical handling of the friable material, and so, even noninvasive methods must be approached with great care. Although low-power x-rays themselves pose no significant danger to inanimate materials, the required transport and handling of the scroll make physical conservation and preservation an ever-present concern. However, once acquired, the volumetric scan data become the basis for all further computations, and the physical object can be returned to the safety of its protective archive.


Segmentation, which is the construction of a geometric model localizing the shape and position of the substrate surface within the scan on which text is presumed to appear, is challenging for several reasons. First, the surface as presented in the scanned volume is not developable, that is, isometric to a plane (1315). Although an isometry could be useful as a constraint in some cases, the skin forming the layers of the En-Gedi scroll has not simply been folded or rolled. Damage to the scroll has deformed the shape of the skin material, which is apparent in the 3D scanned volume, making such a constraint unworkable. Second, the density response of animal skin in the volume is noisy and difficult to localize with methods such as marching cubes (16). Third, layers of the skin that are close together create ambiguities that are difficult to resolve from purely local, shape-based operators. Figure 3 shows four distinct instances where segmentation proves challenging because of the damage and unpredictable variation in the appearance of the surface material in the scan volume.

Fig. 3Segmentation challenges in the En-Gedi scroll, based on examples in the slice view.

Double/split layering and challenging cell structure (left), ambiguous layers with unknown material (middle left), high-density “bubbling” on the secondary layer (middle right), and gap in the primary layer (right).


Our segmentation algorithm applied to the En-Gedi scroll builds a triangulated surface mesh that localizes a coherent section of the animal skin within a defined subvolume through a novel region-growing technique (Fig. 4). The basis for the algorithm is a local estimate of the differential geometry of the animal skin surface using a second-order symmetric tensor and associated feature saliency measures (17). An initial set of seed points propagates through the volume as a connected chain, directed by the local symmetric tensor and constrained by a set of spring forces. The movement of this particle chain through the volume traces out a surface over time. Figure 5 shows how crucially dependent the final result is on an accurate localization of the skin. When the segmented geometry drifts from the skin surface (Fig. 5A), the surface features disappear. When the skin is accurately localized (Fig. 5B), the surface detail, including cracks and ink evidence, becomes visible.

Fig. 4A portion of the segmented surface and how it intersects the volume.

Fig. 5The importance of accurate surface localization on the final generated texture.

(A) Texture generated when the surface is only partially localized. (B) Texture generated when surface is accurately localized.


The user can tune the various parameters of this algorithm locally or globally based on the data set and at any time during the segmentation process. This allows for the continued propagation of the chain without the loss of previously segmented surface information. The segmentation algorithm terminates either at the user’s request or when a specified number of slices have been traversed by all of the particles in the chain.

The global structure of the entire surface is a piecewise composition of many smaller surface sections. Although it is certainly possible to generate a global structure through a single segmentation step, approaching the problem in a piecewise manner allows more accurate localization of individual sections, some of which are very challenging to extract. Although the segmented surface is not constrained to a planar isometry at the segmentation step, the model implicitly favors an approximation of an isometry. Furthermore, the model imposes a point-ordering constraint that prevents sharp discontinuities and self-intersections. The segmented surface, which has been regularized, smoothed, and resampled, becomes the basis for the texturing phase to visualize the surface with the intensities it inherits from its position in the volume.


Once the layers of the scroll have been identified and modeled, the next step is to render readable textures on those layers. Texturing is the assignment of an intensity, “or brightness,” value derived from the volume to each point on a segmented surface. The interpretation of intensity values in the original volumetric scan is maintained through the texturing phase. In the case of micro-CT, intensities are related to density: Brighter values are regions of denser material, and darker values are less dense (18). A coating of ink made from iron gall, for example, would appear bright, indicating a higher density in micro-CT. Our texturing method is similar to the computer graphics approach of “solid texturing,” a procedure that evaluates a function defined over R3 for each point to be textured on the embedded surface (78). In our case, the function over R3 is simply a lookup to reference the value (possibly interpolated) at that precise location in the volume scan.

In an ideal case, where both the scanned volume and localized surface mesh are perfect, a direct mapping of each surface point to its 3D volume position would generate the best possible texture. In practice, however, errors in surface segmentation combined with artifacts in the scan create the need for a filtering approach that can overcome these sources of noise. Therefore, we implement a neighborhood-based directional filtering method, which gives parametric control over the texturing. The texture intensity is calculated from a filter applied to the set of voxels within each surface point’s local neighborhood. The parameters (Fig. 6) include use of the point’s surface normal direction (directional or omnidirectional), the shape and extent of the local texturing neighborhood, and the type of filter applied to the neighborhood. The directional parameter is particularly important when attempting to recover text from dual-sided materials, such as books. In such cases, a single segmented surface can be used to generate both the recto and verso sides of the page. Figure 7 shows how this texturing method improves ink response in the resulting texture when the segmentation does not perfectly intersect the ink position on the substrate in the volumetric scan.

Fig. 6The geometric parameters for directional texturing.

Fig. 7The effect of directional texturing to improve ink response.

(Left) Intersection of the mesh with the volume. (Right) Directional texturing with a neighborhood radius of 7 voxels.



Region-growing in an unstructured volume generates surfaces that are nonplanar. In a scan of rolled-up material, most surface fragments contain high-curvature areas. These surfaces must be flattened to easily view the textures that have been applied to them. The process of flattening is the computation of a 3D to 2D parameterization for a given mesh (61920). One straightforward assumption is that a localized surface cannot self-intersect and represents a coherent piece of substrate that was at one time approximately isometric to a plane. If the writing occurred on a planar surface before it was rolled up, and if the rolling itself induced no elastic deformations in the surface, then damage is the only thing that may have interrupted the isometric nature of the rolling.

We approach parameterization through a physics-based material model (42122). In this model, the mesh is represented as a mass-spring system, where each vertex of the mesh is given a mass and the connections between vertices are treated as springs with associated stiffness coefficients. The mesh is relaxed to a plane through a balanced selection of appropriate forces and parameters. This process mimics the material properties of isometric deformation, which is analogous to the physical act of unwrapping.

A major advantage of a simulation-based approach is the wide range of configurations that are possible under the framework. Parameters and forces can be applied per vertex or per spring. This precise control allows for modeling of not only the geometric properties of a surface but also the physical properties of that surface. For example, materials with higher physical elasticity can be represented as such within the same simulation.

Although this work relies on computing parameterizations solely through this simulation-based method, a hybrid approach that begins with existing parameterization methods [for example, least-squares conformal mapping (LSCM) (23) and angle-based flattening (ABF) (24)] followed by a physics-based model is also workable. The purely geometric approaches of LSCM and ABF produce excellent parameterizations but have no natural way to capture additional constraints arising from the mesh as a physical object. By tracking the physical state of the mesh during parameterization via LSCM or ABF, a secondary correction step using the simulation method could then be applied to account for the mesh’s physical properties.

Merging and visualization

The piecewise nature of the pipeline requires a final merge step. There can be many individually segmented mesh sections that must be reconciled into a composite master view. The shape, location in the volume, and textured appearance of the sections aid in the merge. We take two approaches to the merging step: texture and mesh merging.

Texture merging is the alignment of texture images from small segmentations to generate a composite master view. This process provides valuable user feedback when performed simultaneously with the segmentation process. Texture merging builds a master view that gives quick feedback on the overall progress and quality of segmentation. However, because each section of geometry is flattened independently, the merge produces distortions that are acceptable as an efficiently computed draft view, but must be improved to become a definitive result for the scholarly study of text.

Mesh merging refers to a more precise recalculation of the merge step by using the piecewise meshes to generate a final, high-quality master view. After all segmentation work is complete, individual mesh segmentations are merged in 3D to create a single surface that represents the complete geometry of the segmented scroll. The mesh from this new merged segmentation is then flattened and textured to produce a final master view image. Because mesh merging is computationally expensive compared to texture merging, it is not ideal for the progressive feedback required during segmentation of a scan volume. However, as the performance of algorithms improves and larger segmented surfaces become practical, it is likely that mesh merging will become viable as a user cue during the segmentation process.

Maintaining a provenance chain is an important component of our pipeline. The full set of transformations used to generate a final master view image can be referenced so that every pixel in a final virtually unwrapped master view can be mapped back to the voxel or set of voxels within the volume that contributed to its intensity value. This is important for both the quantitative analysis of the resulting image and the verification of any extracted text. Figure 8demonstrates the ability to select a region and point of interest in the texture image and invert the transformation chain to recover original 3D positions within the volume.

Fig. 8Demonstration of stored provenance chain.

The generated geometric transformations can map a region and point of interest in the master view (left) back to their 3D positions within the volume (right).



Using this pipeline, we have restored and revealed the text on five complete wraps of the animal skin of the En-Gedi scroll, an object that likely will never be physically opened for inspection (Fig. 1). The resulting master image (Fig. 2) enables a complete textual critique, and although such analysis is beyond the scope of this paper, the consensus of our interdisciplinary team is that the virtually unwrapped result equals the best photographic images available in the 21st century. From the master view, one can clearly see the remains of two distinct columns of Hebrew writing that contain legible and countable lines, letters, and words (Fig. 9).

Fig. 9Partial transcription and translation of recovered text.

(Column 1) Lines 5 to 7 from the En-Gedi scroll.


These images reveal the En-Gedi scroll to be the book of Leviticus, which makes it the earliest copy of a Pentateuchal book ever found in a Holy Ark and a significant discovery in biblical archeology (Fig. 10). Without our computational pipeline and the textual analysis it enables, the En-Gedi text would be totally lost for scholarship, and its value would be left unknown.

Fig. 10Timeline placing the En-Gedi scroll within the context of other biblical discoveries.


What is clearly preserved in the master image is part of one sheet of a scripture scroll that contains 35 lines, of which 18 have been preserved and another 17 have been reconstructed. The lines contain 33 to 34 letters and spaces between letters; spaces between the words are indicated but are sometimes minimal. The two columns extracted from the scroll also exhibit an intercolumnar blank space, as well as a large blank space before the first column that is larger than the column of text. This large blank space leaves no doubt that what is preserved is the beginning of a scroll, in this case a Pentateuchal text: the book of Leviticus.

Armed with the extraction of this readable text and its historical context discerned from carbon dating and other related archeological evidence (12), scholars can accurately place the En-Gedi writings in the canonical timeline of biblical text. Radiocarbon results date the scroll to the third or fourth century CE (table S1). Alternatively, a first or second century CE date was suggested on the basis of paleographical considerations by Yardeni (25). Dating the En-Gedi scroll to the third or fourth century CE falls near the end of the period of the biblical Dead Sea Scrolls (third century BCE to second century CE) and several centuries before the medieval biblical fragments found in the Cairo Genizah, which date from the ninth century CE onward (Fig. 10). Hence, the En-Gedi scroll provides an important extension to the evidence of the Dead Sea Scrolls and offers a glimpse into the earliest stages of almost 800 years of near silence in the history of the biblical text.

As may be expected from our knowledge of the development of the Hebrew text, the En-Gedi Hebrew text is not vocalized, there are no indications of verses, and the script resembles other documents from the late Dead Sea Scrolls. The text deciphered thus far is completely identical with the consonantal framework of the medieval text of the Hebrew Bible, traditionally named the Masoretic Text, and which is the text presented in most printed editions of the Hebrew Bible. On the other hand, one to two centuries earlier, the so-called proto-Masoretic text, as reflected in the Judean Desert texts from the first centuries of the Common era, still witnesses some textual fluidity. In addition, the En-Gedi scan revealed columns similar in length to those evidenced among the Dead Sea Scrolls.


Besides illuminating the history of the biblical text, our work on the scroll advances the development of textual imaging. Although previous research has successfully identified text within ancient artifacts, the En-Gedi manuscript represents the first severely damaged, ink-based scroll to be unrolled and identified noninvasively. The recent work of Barfod et al. (26) produced text from within a damaged amulet; however, the text was etched into the amulet’s thin metal surface, which served as a morphological base for the contrast of text. Although challenging, morphological structures provide an additional guide for segmentation that is unlikely to be present with ink-based writing. In the case of the En-Gedi scroll, for instance, the ink sits on the substrate and does not create an additional morphology that can aid the segmentation and rendering process. The amulet work also performed segmentation by constraining the surface to be ruled, and thus developable, to simplify the flattening problem. In addition, segmented strips were assembled showing letterforms, but a complete and merged surface segmentation was not computed, a result of using commercial software rather than implementing a custom software framework.

Samko et al. (27) describe a fully automated approach to segmentation and text extraction of undamaged, scrolled materials. Their results, from known legible manuscripts that can be physically unrolled for visual verification, serve as important test cases to validate their automated segmentation approach. However, the profound damage exhibited in the materials, such as the scroll from En-Gedi, creates its own new challenges—segmentation, texturing, and flattening algorithms—that only our novel framework directly addresses.

The work of Mocella et al. (12) claims that phase-contrast tomography generates contrast at ink boundaries in scans of material from Herculaneum. The hope for phase contrast comes from a progression of volumetric imaging methods (56) and serves as a possible solution to the first step in our pipeline: creating a noninvasive, volumetric scan with some method that shows contrast at ink. Although verifying that ink sits on a page is not enough to allow scholarly study of discovered text, this is an important prelude to subsequent virtual unwrapping. Our complete approach makes such discovery possible.

An overarching concern as this framework becomes widely useful has to do not with technical improvements of the components, which will naturally occur as scientists and engineers innovate over the space of scanning, segmentation, and unwrapping, but rather with the certified provenance of every final texture claim that is produced from a scan volume. An analysis framework must offer the ability for independent researchers to confidently affirm results and verify scholarly claims. Letters, words, and, ultimately, whole texts that are extracted noninvasively from the inaccessible inner windings of an artifact must be subject to independent scrutiny if they are to serve as the basis for the highest level of scholarship. For such scrutiny to occur, accurate and recorded geometry, aligned in the coordinate frame of the original volumetric scan, must be retained for confirmation.

The computational framework we have described provides this ability for a third party to confirm that letterforms in a final output texture are the result of a pipeline of transformations on the original data and not solely an interpretation from an expert who may believe letterforms are visible. Such innate third-party confirmation capability reduces confusion around the resulting textual analyses and engenders confidence in the effectiveness of virtual unwrapping techniques.

The traditional approach of removing a folio from its binding—or unrolling a scroll—and pressing it flat to capture an accurate facsimile obviously will not work on fragile manuscripts that have been burned and crushed into lumps of disintegrating charcoal. The virtual unwrapping that we performed on the En-Gedi scroll proves the effectiveness of our software pipeline as a noninvasive alternative: a technological engine for text discovery in the face of profound damage. The implemented software components, which are necessary stages in the overall process, will continue to improve with every new object and discovery. However, the separable stages themselves, from volumetric scanning to the unwrapping and merging transformations, will become the guiding framework for practitioners seeking to open damaged textual materials. The geometric data passing through the individual stages are amenable to a standard interface through which the software components can interchangeably communicate. Hence, each component is a separable piece that can be individually upgraded as alternative and improved methods are developed. For example, the accurate segmentation of layers within a volume is a crucial part of the pipeline. Segmentation algorithms can be improved by tuning them to the material type (for example, animal skin, papyrus, and bark), the expected layer shape (for example, flat and rolled pages), and the nature of damage (for example, carbonized, burned, and fragmented). The flattening step is another place where improvements will better support user interaction, methods to quantify and visualize errors from flattening, and a comparative analysis between different mapping schemes.

The successful application of our virtual unwrapping pipeline to the En-Gedi scroll represents a confluence of physics, computer science, and the humanities. The technical underpinnings and implemented tools offer a collaborative opportunity between scientists, engineers, and textual scholars, who must remain crucially involved in the process of refining the quality of extracted text. Although more automation in the pipeline is possible, we have now achieved our overarching goal, which is the creation of a new restoration pathway—a way to overcome damage—to reach and retrieve text from the brink of oblivion.


Master view results

The master view image of the En-Gedi scroll was generated using the specific algorithms and processes outlined below. Because we use the volumetric scan as the coordinate frame for all transformations in our pipeline, the resolution of the master view approximately matches that of the scan. The spatial resolution of the volume (18 μm/voxel, isometric) produces an image resolution of approximately 1410 pixels per inch (25,400 μm/inch), which can be considered archival quality. From this, we estimate the surface area of the unwrapped portion to be approximately 94 cm2 (14.57 in2). The average size of letterforms varies between 1.5 and 2 mm, and the pixels of the master view maintain the original dynamic range of 16 bits.

Volumetric scan

Two volumetric scans were performed using a Bruker SkyScan 1176 in vivo Micro-CT machine. It uses a PANalytical microfocus tube and a Princeton Instruments camera. With a maximum spatial resolution of 9 μm per voxel, it more than exceeded the resolution requirements for the En-Gedi scroll. A spatial resolution of 18 μm was used for all En-Gedi scans. Additionally, because this is an in vivo machine, the scroll could simply be placed within the scan chamber and did not need to be mounted for scanning. This limited the risk of physical damage to the object.

An initial, single field-of-view scan was done on the scroll to verify the scan parameters and to confirm that the resolution requirements had been met. This scan was performed at 50 kV, 500 μA, and 350-ms exposure time, with added filtration (0.2 Al) to improve image quality by absorbing lower-energy x-rays that tend to produce scattering. The reconstructions showed very clear separation of layers within the scroll, which indicated a good opportunity for segmentation. The scan protocol was then modified to increase contrast, where the team suspected that there may be visible ink. To acquire data from as much of the scroll as possible, the second and final scan was an offset scan using four connected scans. Final exposure parameters of 45 kV, 555 μA, and 230 ms were selected for this scan. The data were reconstructed using Bruker SkyScan’s NRecon engine, and the reconstructed slices were saved as 16-bit TIFF images for further analysis.


The basis for the algorithm is a local estimate of the differential geometry of the animal skin surface using a second-order symmetric tensor and associated feature saliency measures (17). The tensor-based saliency measures are available at each point in the volume. The 3D structure tensor for point p is calculated asEmbedded Image(1)where ∇u(p) is the 3D gradient at point pgσ is a 3D Gaussian kernel, and denotes convolution. The eigenvalues and eigenvectors of this tensor provide an estimate of the surface normal at pEmbedded Image(2)

The algorithm begins with an ordered chain of seed points localized to a single slice by a user. From the starting point, each particle in the chain undergoes a force calculation that propagates the chain forward through the volume. This region-growing algorithm estimates a new positional offset for each particle in the chain based on the contribution of three forces: gravity (a bias in the direction of the long axis of the scroll), neighbor separation, and the saliency measure of the structure tensor. First, n(p) is biased along the long axis of the scroll by finding the vector rejection of an axis-aligned unit vector Z onto n(p)Embedded Image(3)

To keep particles moving together, an additional restoring spring force S is computed using a spring stiffness coefficient k and the elongation factors Xl and Xr between the particle and both its left and right neighbors in the chainEmbedded Image(4)

In the final formulation, a scaling factor α is applied to G, and the final positional offset for the particle is the normalized summation of all termsEmbedded Image(5)

The intuition behind this framework is the following: The structure tensor estimates a surface normal, which gives a hint at how a layer is moving locally through points (Eqs. 1 and 2). The layer should extend in a direction that is approximated locally by its tangent plane: the surface normal. The gravity vector nudges points along the major axis around which the surface is rolled, which is a big hint about the general direction to pursue to extend a surface. Moving outward, away from the major axis and across layers, generally defeats the goal of following the same layer. The spring forces help to maintain a constant spacing between points, restraining them from moving independently. These forces must all be balanced relative to one another, which is done by trial and error.

We applied the computed offset iteratively to each particle, resulting in a surface mesh sampled at a specific resolution relative to the time step of the particle system. The user can tune the various parameters of this algorithm based on the data set and at any time during the segmentation process. We also provided a feedback interface whereby a human user can reliably identify a failed segmentation and correct for mistakes in the segmentation process.

For each small segmentation, we used spring force constants of 0.5 and an α scaling factor of 0.3. Segmentation chain points had an initial separation of 1 voxel. This generally produced about six triangles per 100 μm in the segmented surface models. When particles crossed surface boundaries because the structure tensor did not provide a valid estimate of the local surface normal, the chain was manually corrected by the user.


Two shapes were tested for the shape of the texturing neighborhood: a spheroid and a line. The line shape includes only those voxels that directly intersect along the surface normal. When the surface normal is accurate and smoothly varying, the line neighborhood allows parametric control for the texture calculation to incorporate voxels that are near but not on (or within) the surface. The line neighborhood leads to faster processing times and less blurring on the surface, although the spheroid neighborhood supports more generalized experimentation—the line neighborhood is a degenerate spheroid. We settled on bidirectional neighborhoods (voxels in both the positive and negative direction) using a line neighborhood with a primary axis length of 7 voxels. We filtered the neighborhood using a Max filter because the average ink response (density) in the volume was much brighter than the ink response of the animal skin.


Our flattening implementation makes use of Bullet Physics Library’s soft body simulation, which uses position-based dynamics to solve the soft body system. Points along one of the long edges of the segmentation were pinned in place while a gravity force along the x axis was applied to the rest of the body. This roughly “unfurled” the wrapping and aligned the mesh with the xz plane. A gravity force along the y axis was then applied to the entire body, which pushed the mesh against a collision plane parallel to the xz plane. This action flattened the curvature of the mesh against the plane. A final expansion and relaxation step was applied to smooth out wrinkles in the internal mesh.


In total, around 140 small segmentations were generated during the segmentation process. These segmentations were then mesh-merged to produce seven larger segmentations, a little over one for each wrap of the scroll. Each large segmentation was then flattened and textured individually. The final set of seven texture images was then texture-merged to produce the final master view imagery for this paper. All merging steps for this work were performed by hand. Texture merges were performed in Adobe Photoshop, and mesh merging was performed in MeshLab. An enhancement curve was uniformly applied to the merged master view image to enhance visual contrast between the text and substrate.


Supplementary material for this article is available at

table S1. Radiocarbon dating results of the En-Gedi scroll (25).

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank D. Merkel (Merkel Technologies Ltd.) who donated the volumetric scan to the IAA. Special thanks to the excavators of the En-Gedi site D. Barag and E. Netzer (Institute of Archaeology at the Hebrew University of Jerusalem). Radiocarbon determination was made at the DANGOOR Research Accelerator Mass Spectrometer (D-REAMS) at the Weizmann Institute. We thank G. Bearman (imaging technology consultant, IAA.) for support and encouragement. W.B.S. acknowledges the invaluable professional contributions of C. Chapman in the editorial preparation of this manuscript. Funding: W.B.S. acknowledges funding from the NSF (awards IIS-0535003 and IIS-1422039). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF. W.B.S. acknowledges funding from Google and support from S. Crossan (Founding Director of the Google Cultural Institute).Author contributions: W.B.S. conceived and designed the virtual unwrapping research program. C.S.P. directed the software implementation team and assembled the final master view. P.S. acquired the volumetric scan of the scroll and its radiocarbon dating and provided initial textual analysis. M.S. and E.T. edited the visible text and interpreted its significance in the context of biblical scholarship. Y.P. excavated the En-Gedi scroll on May 5, 1970. The manuscript was prepared and submitted by W.B.S. with contributions from all authors.Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors. All scan data and results from this paper are archived at the Department of Computer Science, University of Kentucky (Lexington, KY) and are available at

View Abstract

Battery technology set to break barriers

Battery technology set to break barriers

Wind farm
Once renewable energy can be stored for use on demand, Britain could become self-sufficient in its energy usage CREDIT: CHARLOTTE GRAHAM/REX SHUTTERSTOCK 

The world’s next energy revolution is probably no more than five or ten years away. Cutting-edge research into cheap and clean forms of electricity storage is moving so fast that we may never again need to build 20th Century power plants in this country, let alone a nuclear white elephant such as Hinkley Point.

The US Energy Department is funding 75 projects developing electricity storage, mobilizing teams of scientists at Harvard, MIT, Stanford, and the elite Lawrence Livermore and Oak Ridge labs in a bid for what it calls the ‘Holy Grail’ of energy policy.

You can track what they are doing at the Advanced Research Projects Agency-Energy (ARPA-E). There are plans for hydrogen bromide, or zinc-air batteries, or storage in molten glass, or next-generation flywheels, many claiming “drastic improvements” that can slash storage costs by 80pc to 90pc and reach the magical figure of $100 per kilowatt hour in relatively short order.

“Storage is a huge deal,” says Ernest Moniz, the US Energy Secretary and himself a nuclear physicist. He is now confident that the US grid and power system will be completely “decarbonised” by the middle of the century.

The technology is poised to overcome the curse of ‘intermittency’ that has long bedevilled wind and solar. Surges of excess power will be stored for use later at times when the sun sets, and consumption peaks in the early evening.

This transforms the calculus of energy policy. The question for the British government as it designs a strategy fit for the 21st Century – and wrestles with an exorbitant commitment to Hinkley Point – is no longer whether this form of back-up power will ever be commercially viable, but whether the inflection point arrives in the early-2020s or in the late 2020s.

One front-runner – a Washington favourite – is an organic flow battery at Harvard that uses quinones from cheap and abundant sources such as rhubarb or oil waste. It is much cheaper and less toxic than current flow batteries based on the rare metal vanadium. Its reactions are 1,000 times faster.

Harvard professor Michael Aziz working on his revolutionary ‘rhubarb battery’ CREDIT: HARVARD

Professor Michael Aziz, leader of the Harvard project, said there are still problems to sort out with the “calendar life” of storage chemicals but the basic design is essentially proven.

“We have a fighting chance of bringing down the capital cost to $100 a kilowatt hour, and that will change the world. It could complement wind and solar on a very large scale,” he told the Daily Telegraph.

The latest refinement is to replace toxic bromine with harmless ferrocyanide – used in food additives. The battery cannot catch fire. It is safe and clean. “This is chemistry I’d be happy to put in my basement,” he said.

The design is delightfully simple. It uses a tank of water. You could have one at home in Los Angeles, Lagos, Buenos Aires, Delhi, or Guangzhou, storing solar power in the day to drive your air-conditioning at night. It could be scaled up for a 500 megawatt wind farm.

The Harvard organic flow battery. It can run off rhubarb CREDIT: US ENERGY DEPARTMENT

Italy’s Green Energy Storage has the European licence. It is building a prototype with the Kessler Foundation at Trento University, backed by EU funds. “We have a chemistry that is ten times cheaper than anything on the market,” said Salvatore Pinto, the chairman.

“We are speaking to three utilities in Europe and we will install our first battery as a field test next year,” he said.

It is a race. Tim Grejtak, an energy expert at Lux Research, said Lockheed Martin and Pacific Northwest labs are both working on their own organic flow batteries, while others are developing variants with designed molecules.

I do not wish to single out this particular technology. I cite it as an example of how fast the picture is evolving as the world’s scientific superpower mobilizes in earnest, and investors start to chase the immense prize. Consultants Mckinsey estimate that the energy storage market will grow a hundredfold to $90bn a year by 2025.

Once storage costs approach $100 per kilowatt hour, there ceases to be much point in building costly ‘baseload’ power plants such as Hinkley Point. Nuclear reactors cannot be switched on and off as need demands – unlike gas plants. They are useless as a back-up for the decentralized grid of the future, when wind, solar, hydro, and other renewables will dominate the power supply.

I will be writing about the economics of offshore wind in coming days but bear in mind that renewables generated 18pc of UK power last year, and this is expected to double by the late 2020s as wind and solar capacity reach 50 gigawatts (GW). Once the power can be stored for overnight use, there will be extended periods in the summer when no base-load is needed whatsoever.

Perhaps the Hinkley project still made sense in 2013 before the collapse in global energy prices and before the latest leap forward in renewable technology. It is madness today.

The latest report by the National Audit Office shows that the estimated subsidy for these two reactors has already jumped from £6bn to near £30bn. Hinkley Point locks Britain into a strike price of £92.50 per megawatt hour – adjusted for inflation, already £97 – and that is guaranteed for 35 years.

That is double the current market price of electricity. The NAO’s figures show that solar will be nearer £60 per megawatt hour by 2025. Dong Energy has already agreed to an offshore wind contract in Holland at less than £75.

Michael Liebreich from Bloomberg New Energy Finance says the Hinkley Point saga will be taught for generations as a case study in how not to run a procurement process. “The obvious question is why this train-wreck of a project was not killed long ago,” he said.

All the extra power capacity added since the mid-1990s in Europe has been renewableCREDIT: EWEA

Theresa May has inherited a poisonous dossier, left with the invidious choice of either offending China or persisting with a venture that no longer makes any economic sense. She may have to offend China – as tactfully as possible, let us hope –  for the scale of the folly has become crushingly obvious.

Every big decision on energy strategy by the British government or any other government must henceforth be based on the working premise that cheap energy storage will soon be a reality.

This country can achieve total self-sufficiency in power at viable cost from our own sun, wind, and waters within a generation. Once we shift to electric vehicles as well, we will no longer need to import much oil either. Rejoice.


Autodesk Just Gave Every Fab Lab Access to $25,000 in Design Software

Autodesk Just Gave Every Fab Lab Access to $25,000 in Design Software


Last year Autodesk granted Fab Labs, a global network of workshops and makerspaces connected through the Fab Foundation, access to Tinkercad, 123D Circuits, and Fusion 360 Labs, but that was just the beginning. Earlier this year, Autodesk expanded their generosity to the Fab Foundation’s Fab Academy (which is the foundation’s training initiative) by offering them access to their entire Product Design Collection, and today Autodesk announced that they will be offering all registered Fab Labs in the Fab Foundation’s network 10 licenses to the Product Design Collection.


The Autodesk suite of software is a reliable favorite of professionals and hobbyists alike, and having access to these tools is invaluable to a student or part-time inventor who might not otherwise be able to get their ideas off the ground.

The Product Design Collection includes the 14 software programs and would cost a person $2,460 a year to buy all of them as a collection, but individually purchasing the software would be even more expensive (Inventor Professional alone would be $1,890 and AutoCAD another $1,680). This means that each Fab Lab stands to gain $24,600 worth of software and that (given the fact the network includes over 1000 Fab Labs) there’s $24,600,000 in software available to the Fab Foundation’s members.

This offer is part of Autodesk’s extensive efforts to support the global maker movement through free software to students, educators, non-profits and educational institutions (like the Fab Foundation), theirEntrepreneur Impact Program for start-ups and hobbyists, and by open-sourcing design files for our 3D printer Ember.

Speaking about their partnership with the Fab Foundation specifically, Autodesk’s Rama Dunayevich, Senior Manager of Brand Partnerships, speaking about the partnership with the Fab Foundation specifically “At each step of the way, we are learning more about how best to support Fab Labs so we can improve and deepen our relationship. We are using an initial one-year renewable subscription term to collect feedback and data to be in a better position to consider further expansions, refinements and customization based on that data.”

The details of Autodesk’s offer are as follows:

Participating Fab Labs will receive a grant that includes 10 subscriptions to the Autodesk Product Design Collection for the Fab Lab which can be used on or off-site by its members. The subscription includes the professional-grade 3D design tools needed to succeed in the changing landscape for product makers, hobbyists and manufacturers including Fusion 360, Inventor Professional, AutoCAD, ReCap 360 Pro, 3ds Max, 25 gigabytes of cloud storage and more.

The Product Design Collection includes:

·         Inventor Professional

·         AutoCAD

·         AutoCAD Architecture

·         AutoCAD Electrical

·         AutoCAD Mechanical

·         AutoCAD 360 Pro

·         Cloud storage (25 GB)

·         Factory Design Utilities

·         Fusion 360

·         Navisworks Manage

·         ReCap 360 Pro

·         Rendering in A360

·         Vault Basic

·         3ds Max

Fab Labs registered as part of the Fab Foundation’s network of global Fab Labs can apply for the grant now. There is no limit to the number of grants Autodesk is willing to fulfill and no additional requirements besides being a registered Fab Lab.

To become a registered Fab Lab your space must have 5 different types of fabrication tools (CNC machining, 3D Printing, microelectronics workstations, etc.); it must be free and open to the public (for at least some portion of time if operated out of a private school or university); abide by the Fab Lab Charter; and participate in your community.  You can register here. There is no licensing fee and someone from your region will confirm that you meet the qualifications.

This announcement is being made now to coincide with the Fab 12 conference being held in Shenzhen. This conference is a gathering of Fab Labs members to discuss the future of community led, education-focused workspaces and participate in workshops centered on different fabrication projects.

Muslim doctor: My patient refused to let me treat her because of my religion



Muslim doctor: My patient refused to let me treat her because of my religion

 August 10

Making my rounds in the hospital one day, I put my stethoscope to a patient’s chest while she kept her eyes fixed on the television screen over my shoulder.

Hours before, bombs had torn through an airport and a train station in Brussels. My 65-year-old patient watched a flurry of images on Fox News showing unfathomable carnage, and I went through the all-too-familiar ritual of hoping that the perpetrators would not be identified as Muslim, that members of my faith would not be considered guilty by inexplicable association.

The sounds of my patient’s voice rose, eclipsing the thump of her heartbeat that I was painstakingly trying to hear.

She sounded distressed, anguished even, about the loss of the innocent lives on the TV screen. “These foreign people only come here to kill and ruin things,” she said. Then she said Donald Trump is right: America should ban all Muslims from immigrating here.

And then perhaps she noticed the subtle change in my facial expression. “I’m sorry, but your people and people who look like you make me uncomfortable,“ she said.

She refused to let me treat her.

I stood aghast at the bedside, wondering how my humanity and years of medical training had been negated by the acts of a sinister few an ocean away. With her words, the ascendant xenophobia of our time infiltrated the sacred patient-doctor relationship.

I had understood, in the abstract, the threat of Trump’s demagoguery and petulance. But until that moment, the bile he spewed seemed confined to Twitter and to rallies in faraway places. I didn’t think it would ever reach me, a physician, born here in the United States.

I should have known better.

Medicine is not practiced in a vacuum. It’s subject to the same influences affecting the rest of our society. In our current political environment, the toxic, Islamophobic rhetoric intended to incite and galvanize voters is of course seeping into hospitals and clinics.

A study published last year in AJOB Empirical Bioethics of Muslim doctors, who comprise 5 percent of U.S. physicians, found that 1 in 10 of these doctors has had a patient refuse their medical care because they are Muslim. Clearly my experience was not isolated.

That day, I deferred to my medical students to hastily complete the patient’s physical exam while I stepped aside and observed from the periphery. My impulses wrestled with my discretion, trying to suppress my desire to respond. Anger would only further alienate the patient, I knew.

I wondered what my responsibilities to the patient were in the face of this bigotry. Did any remain?

As a physician, I thought first of the Hippocratic oath, a doctor’s steadfast commitment to a patient. We doctors heal the affluent or the dispossessed. We heal regardless of sexual orientation, gender, religion or race. This oath compels us to invariably protect the patient from our biases and darkest demons. But it offers little wisdom when the physician is the one subjected to intolerance.

I needed more. And as I attempted to reconcile my desire to preserve my wounded personal dignity with the principles enshrined in the Hippocratic oath, I found my Islam.

Clarity arrived as I remembered one of my favorite verses from the Koran: “Believers, stand firm for God, be witnesses for justice. Never allow the hatred of people to prevent you from being just. Be just, for this is closest to righteousness.”

The concept of justice captured in this verse animated Muslims during the nascent stages of Islam. The Prophet Muhammad’s unprecedented message of egalitarianism threatened the entrenched social order and thus provoked considerable malice at the time. Yet this early Muslim community endured the ridicule and persecution because it was inspired by something far greater than itself.

Retreating into cynicism and rage was too tempting in the aftermath of my experience with the patient, but it would be an affront to the legacy and sacrifice of those that came before me.

In the end, this was all part of my personal inner jihad, which literally means “a struggle.” For a Muslim, jihad means the struggle of the soul to topple the barriers which prevent the realization of a divinely inspired life, a life that promotes the virtues of compassion, understanding and justice. A selfish focus on my own damaged ego detracted from this purpose.

My decision to work as a physician in the public setting of a hospital meant that I was an ambassador for my faith, whether I wanted to be or not. That’s an important role: the Pew Research Center has found that Americans are more likely to have a favorable opinion of Muslims if they know one.

As I reflected on the conversation with my patient that had gone awry, I realized the value in the seemingly mundane conversations I share daily with other patients. Perhaps the exchanges about their lives and mine are transformative. Our interactions allow us to see our common hopes and fears in each other.

Trump and others of his ilk have sought to sow division by stripping groups like Muslims, African Americans, Mexicans and women of their common humanity. His candidacy has brought long-hidden rancor into the open, even into our hospitals.

The finest and noblest traditions of medicine, however, are transcendent. My personal faith subsumes the Hippocratic oath and preaches mutual respect and tolerance. These are the sources of my continued dignity.

And they allowed me to knock on the same patient’s door again the next day.

Jalal Baig is an oncologist and a writer in Chicago.

Chinese Hui Muslims

China’s Complex Relationship With Islam Is Reflected in Ties to Hui

 How Islam Coexists With Communism in Atheist China 5:41

YINCHUAN, China — This imam wears heels — and a gold-flecked headscarf.

Ma Lijuan is one of the faces of state-backed Islam in officially atheist Communist China.

Female imams like her are unique to China’s Hui, a Muslim minority group — as are their women-only mosques.

Ma hails from the Ningxia Hui Autonomous Region, a sparsely populated province straddling an arid plateau on the Yellow River basin in north-central China.

“Here, [religion] is very normal for us,” said Ma, 33. “Our constitution grants us the freedom of religious belief, and we are thankful to live here.”

Image: Ma Lijuan
Ma Lijuan, an imam at Nanguan mosque in central Yinchuan, teaches Quran classes to dozens of women who repeat her chants in Arabic. David Lom / NBC News

Ma, who is married with two boys, helps lead a congregation of hundreds at the Nanguan Mosque in the provincial capital of Yinchuan. The mosque — with its onion-shaped domes and treed courtyard — dates to the 1900s but was refurbished in the 1980s.

“As a teacher, spreading Islamic truth is my duty,” Ma told NBC News. “When we are educating a Muslim woman we are also educating a family.”

Every morning Ma teaches Quran classes to dozens of mostly middle-aged and older women who repeat her chants in Arabic. She speaks in a gentle yet confident voice, and her infectious laugh punctuates conversations.

Ethnically Chinese Hui Muslims speak Mandarin and in recent decades have nurtured a coexistence with the Communist Party. The party has in return rewarded the community — which traces its roots to 10th-century Arab traders who married Chinese women — with the latitude and support to establish mosques, religious schools and museums.

However, the two haven’t always had such a harmonious relationship.

During the Cultural Revolution, religious communities throughout China were persecuted, and the Hui were no exception. After the country emerged from the tumult in 1976, the Hui reinstated the practice of their faith.

 Former Red Guard Remembers Communist Revolution 1:53

China’s constitution provides for freedom of religion, but within very strict limits. Five “recognized religions” — including Christianity and Islam — are permitted to worship at state-approved sites. Proselytizing in public is illegal and unregistered groups, such as Tibetan Buddhists or Falun Gong, routinely encounter problems.

Not all of China’s 23 million Muslims, a legacy of traders who centuries ago ventured from the Middle East along the Silk Road, practice as freely as the Hui do.

Turkic-speaking Uighurs have increasingly chafed at large-scale immigration of China’s dominant Han Chinese. Ethnic tensions fueled violence that Chinese authorities point to as justification for religious restrictions and repressive tactics.

Image: A veiled Muslim Uighur woman in Kashgar in 2014
A veiled Muslim Uighur woman walks passed a statue of Mao Zedong in Kashgar in China’s restive Xinjiang Province in 2014. Kevin Frayer / Getty Images

This year the government barred Uighur Muslim civil servants, students and children in the far western region of Xinjiang from fasting for the Muslim holy month of Ramadan.

The rest of the year, burqas, long beards and calls to prayer over loudspeakers are banned throughout Xinjiang. People under the age of 18 are not allowed to enter mosques at all.

The government’s unease with Islam intensified with the rise of Islamic extremism and as groups like ISIS recruited followers in China. Late last year, a digital recording of an ISIS chant in Mandarin calling on Muslims to “wake up” and “take up weapons to fight” was posted online.

“It would be naïve to say China does not face a problem with jihadi-style extremism,” said James Leibold, an expert on China’s ethnic minorities at Australia’s LaTrobe University. “But I don’t think that problem is nearly on the scale that the government perceives it to be. They have overblown it, which is dangerous and counter-productive because it whips up anti-Islamic fear and suspicion.”

Gallery: Uighurs in Kashgar: China’s Muslims Try To Preserve Traditions

Aware of how its crackdown in Xinjiang might affect its image in the Middle East, China recently issued a white paper on religion that declared “no Xinjiang citizen has been punished because of his or her rightful religious belief.”

Beijing, meanwhile, has invested heavily in Yinchuan and the surrounding area.

The makeover includes a new airport terminal to handle what local officials hope will accommodate a flood of Muslim tourists from the Middle East as China seeks to boost its image in and trade with the Arab world. The government is also promoting Hui-made products and real estate developments as part of its ambitious “One Belt, One Road” strategy to revitalize trade along the ancient Silk Road.


Image: Hui men in the courtyard of Najiahu mosque
Hui men in the courtyard of Najiahu Mosque. Janis Mackey Frayer / NBC News

Even street signs are being translated to Arabic.

“Unlike some other minorities, Hui people follow the leadership of the government,” said singer Ma Shouyun. “The state supports our religion very much.”

Because of all that the Hui are allowed to practice in relative peace.

On a recent evening as the call to prayer echoed across a courtyard, dozens of men adjusted their white caps under the shade of a locust tree and filed into Najiahu Mosque. It is a ritual practiced five times a day by Muslims all over the world.

But even in Yinchuan amidst state support most Hui remain conscious of their delicate position. When the conversation with Ma, the singer, shifted from Hui culture to Islam, a nearby shopkeeper interrupted him.

“No questions about religion,” she warned him, sparking a tense discussion with bystanders about the risk of irritating officials.

Few of them noticed the singer take off his TV microphone, straighten his cap and quietly walk away. 

Japanese Supreme Court Says it is OK to Profile Muslims

Top court green-lights surveillance of Japan’s Muslims

Legal challenge to police profiling of North Asian country’s Islamic population dismissed by Supreme Court.

Listen to this page using ReadSpeaker
Junko Hayashi is the lawyer for plaintiffs who challenged police surveillance of Japan's Muslim community [Ian Munroe/Al Jazeera]
Junko Hayashi is the lawyer for plaintiffs who challenged police surveillance of Japan’s Muslim community [Ian Munroe/Al Jazeera]

Tokyo, Japan – Mohamed Fujita used to host religious study groups at his home that were open to all Muslims. But today he’s afraid to invite strangers, in case they’re police informants.

Extensive surveillance has put many people of his faith on edge, he says, sowing mistrust.

A native of Japan who converted to Islam more than two decades ago, Fujita was one of 17 plaintiffs in a lawsuit that challenged blanket monitoring of the country’s followers of Islam. His name has been changed in this story to protect his identity, after police documents labelling him a possible security threat were leaked online.

“They made us terrorist suspects,” he says. “We never did anything wrong – on the contrary.”

Islam in the Land of the Rising Sun

Fujita’s wife first noticed the couple was being followed by law enforcement in the early 2000s. He says he would go out of his way to cooperate with officers when they would occasionally approach him. But they eventually asked that he report on other members of his mosque and he refused.

Then came the leak in 2010 of 114 police files, which revealed religious profiling of Muslims across Japan. The documents included resumé-like pages listing a host of personal information, including an individual’s name, physical description, personal relationships and the mosque they attended, along with a section titled “suspicions”.

A skyline view of a mosque in Japan’s capital, Tokyo [Ian Munroe/Al Jazeera] 

The files also showed by the time the 2008 G8 summit was held in Hokkaido, northern Japan, at least 72,000 residents from Organisation of Islamic Conference countries had been profiled – including about 1,600 public school students in and around Tokyo.

Police in the capital had also been surveilling places of worship, halal restaurants, and “Islam-related” organisations, the documents showed.

Within a few weeks of the leak, the data had been downloaded from a file-sharing website more than 10,000 times in more than 20 countries.

Fujita and the other plaintiffs, many of whom were originally from Middle Eastern or North African countries, sued in the hope the courts would deem the police practices illegal. Their lawyers said police had violated their constitutional rights to privacy, equal treatment, and religious freedom.

After two appeals, the Supreme Court dismissed the case on May 31.

The justices concurred with a lower court that the plaintiffs deserved a total of ¥90 million ($880,000) in compensation because the leak violated their privacy. But they did not weigh in on the police profiling or surveillance practices, which a lower court ruling had upheld as “necessary and inevitable” to guard against the threat of international terrorism.

“We were told we don’t have a constitutional case,” says Junko Hayashi, a lawyer for the plaintiffs. “We’re still trying to figure out, how is it not constitutional?”

Law enforcement mostly ignored the case. One of the few public statements they made came at a United Nations human rights committee hearing on the matter in 2014. An official from the National Police Agency said “details of information-gathering activities to prevent future terrorism could not be disclosed”, but that “police collected information according to the law”, according to UN records.

Japan’s rich Muslim past and present

Some have defended the surveillance of Muslims, including Naofumi Miyasaka, a professor at the National Defense Academy of Japan. He describes the data leak as “the biggest failure in the history of Japan’s counterterrorism” because it would have hurt the ability of law enforcement to gather intelligence on potential threats through “mutual trust and cooperation between police and informants”.

The Supreme Court decision generated few headlines and little public debate in Japan. Local media outlets had covered the legal proceedings by focusing on the leak of information, tiptoeing around the police surveillance issue.

The most prominent public figure to comment on the Supreme Court decision was NSA whistleblower Edward Snowden, who spoke via video linkup at a symposium on government surveillance in Tokyo on June 4.

Former NSA agent Edward Snowden discussed surveillance of Japanese Muslims earlier this month [Ian Munroe/Al Jazeera]

“People of the Islamic faith are more likely to be targeted … despite not having any criminal activities or associations or anything like that in their background, simply because people are afraid,” said Snowden, who once worked in western Tokyo at a liaison facility between American and Japanese intelligence services.

“But in Japan, let’s look seriously at that. The Aum Shinrikyo was the last significant terrorist event in Japan, and that was over 20 years ago,” he added, referring to the 1995 sarin gas attack on a Tokyo subway that killed 13 people and injured 6,000 others.

“This wasn’t a fundamental Islamic extremist group, this was a crazy doomsday cult that wanted to make their founder the new emperor of Japan.”

Other observers have also questioned whether monitoring a particular religious group en masse is an effective counterterrorism strategy.

“Germany had a similar programme, and they investigated and maintained files on 30,000 Muslims in Germany and didn’t find, apparently, a single terrorist among them,” says Jeff Kingston, director of Asian studies at Temple University in Tokyo. “So it’s not clear that this indiscriminate surveillance is useful.”

Hiroshi Miyashita, a law professor at Chuo University who’s an expert on privacy issues, says he believes the lawsuit was the first major legal case in Japan to focus on mass surveillance. However, he adds that a state secrets law that came into force in 2014 would shield the issue from scrutiny by the public and the courts in future.

“Even judges cannot access information” about police practices under the new law, he says.

Japan’s newest and largest mosque opens its doors

The Tokyo Metropolitan Police and the National Police Agency declined a request to comment on the court decision, and would not confirm whether they continue to profile and monitor Japan’s Muslim community.

But the plaintiffs’ lawyer, Hayashi, who is Muslim, says she believes the surveillance has only intensified.

“It’s a really, really difficult thing to deal with, especially for the kids growing up here,” she adds. “The police have been dealing with them as future terrorists.”

While the lawsuit wasn’t successful, Fujita says he has learned from the experience. To him, the top court’s unwillingness to weigh their constitutional arguments shows that Japan’s judiciary isn’t an independent branch of government.

And years after the police surveillance came to light, he says it continues to rob many Muslims of a sense of trust, which he describes as “the foundation of human relationships”.

A Palestinian Village Tries to Protect a Terraced Ancient Wonder of Agriculture


Palestinian farmers in Battir, a West Bank village near Bethlehem, use a Roman-era irrigation system to water their crops. Credit Abir Sultan/European Pressphoto Agency

BATTIR, West Bank — In this scenic Palestinian village in the West Bank hills near Bethlehem, just south of Jerusalem, a week is said to last eight days, not seven. That is because Battir’s eight extended families take daily turns watering their crops from the natural springs that feed their ancient agricultural terraces, a practice they say has worked for centuries.

The water flows through a Roman-era irrigation system down into a deep valley where a railway track — a section of the Jaffa-Jerusalem railway built in Ottoman times — roughly marks the 1949 armistice line between the West Bank and Israel. The area is dotted with tombs and ruins upon ruins of bygone civilizations.

When the World Heritage Committee of Unesco — the United Nations Educational, Scientific and Cultural Organization — meets in St. Petersburg, Russia, over the next two weeks, this pastoral area will be thrust into the spotlight at least momentarily as the villagers and conservation experts fight to save what they say is a unique living cultural and historical landscape.

The experts say the Battir terraces are under imminent threat because Israel plans to build a section of its West Bank security barrier right through the valley, parallel to the railway track. They are seeking to have Battir nominated as a World Heritage site on an emergency basis, a move that might persuade Israel to change its plans for the construction.

“The people here constructed their village while always preserving the terraces,” said Hassan Muamer, 27, a civil engineer working for the Battir Landscape Eco-Museum. “It was part of the mentality,” he added. “It is living history.”

But the effort to secure a nomination for Battir has been bogged down by internal Palestinian disagreements, designs and interests. The formal submission of the case was blocked at the last minute on the grounds that it had come too late. Instead, the Palestinian delegation to Unesco is pushing a higher-profile, more political effort to have Bethlehem’s venerated Church of the Nativity and pilgrimage route inscribed on the list of World Heritage sites on an emergency basis.

A panel of experts has already determined that although the church needs renovation and conservation, it does not appear to be in imminent danger and therefore does not qualify for emergency status. Leaders of the three churches that share control of the Church of the Nativity, always leery of prospective changes to the delicate status quo, also expressed some early reservations.

Continue reading the main story



When Unesco granted Palestinians full membership in the organization last October, Israel and the United States viewed the development as part of a contentious, wider Palestinian campaign for international recognition of statehood in the absence of an agreement with Israel. The step cost Unesco one-quarter of its yearly budget — 22 percent, or about $70 million, contributed by the United States, and 3 percent contributed by Israel.

Now some Palestinian and Western officials say that by pushing the case of the Church of the Nativity, the traditional birthplace of Jesus, the Palestinian leadership is putting prestige above professional and technical considerations.

In response to the criticism, the Palestinian ambassador to Unesco, Elias Sanbar, wrote a letter condemning what he called “a persistent campaign of rumors aimed at discrediting Bethlehem’s candidacy” by “those who do not want to see Palestine exercise its legitimate rights.” He attached a statement from two of the three church leaders expressing their thanks to the Palestinian leadership for its efforts to safeguard and advance the Christian congregations’ freedom and cause.


Battir is seeking World Heritage status from Unesco. Credit The New York Times

Still, experts in the Palestinian territories say Battir is in more urgent need of protection.

“If Battir is submitted only next year, it may be too late,” said Giovanni Fontana Antonelli, the cultural heritage program specialist at the Unesco office in Ramallah, in the West Bank. “If the wall goes through the valley, it will totally destroy the integrity of the site,” he added.

Noting that the terraces are supported by dry stone walls made up of many millions of stones, Mr. Fontana characterized the valley as “not monumental but historical, an example of outstanding engineering.”

“The work of human beings there needs to be valued,” he said. “It is the work of centuries.”

Israel says its barrier, a system of fences and walls, razor wire and patrol roads, is essential to prevent Palestinian suicide bombers from reaching Israeli cities.

The villagers have petitioned the Supreme Court in Israel to have the barrier rerouted here to prevent the destruction of the striking beauty of the area and its ancient system of cultivation. A court decision is pending. The conservationists hope that a recommendation from the World Heritage Committee may help persuade the court not to reject the villagers’ petition.

Local Palestinians like Raed Samara, a planning and development expert who has been active in promoting the case of Battir, say construction of a barrier would destroy the tranquillity that has prevailed here for decades.

The steep slopes across from Battir are in Israel, making this shared landscape a transboundary site in the Unesco lexicon.

“Nobody thinks that Israel’s security concerns are not legitimate or important,” said Gidon Bromberg, the Israeli director of Friends of the Earth Middle East, an organization that works to promote cooperation on environmental issues in Israel, Jordan and the Palestinian territories. But, he added, “there are alternative ways to bring about security without destroying 4,000 years of cultural heritage for the Israelis, the Palestinians and all of humanity.”

On a recent evening, Mohannad Abu Hassan, a schoolteacher, was working a small triangular plot in the valley with his son Muhammad, 12. Water poured in from one corner as they turned the rich soil planted with green beans, zucchini, eggplant and chard. As soon as they were finished watering, a sprightly elderly woman, a distant relative, skipped down to a nearby plot across the railway track and turned her water on. In the old core of the village, children bathed in the cool waters of the central spring.

Until the late 1940s, Battir was the last stop before Jerusalem on the Jaffa-Jerusalem railway. The train platform used to turn into a bustling market, and the villagers maintained strong connections with the city. The train does not stop here anymore, and most of the produce is now for home use or for local sale. But the villagers are keeping up with the times, swapping news about the Unesco effort through a Facebook group of 2,000 residents and supporters.

Akram Bader, the mayor of Battir, recently traveled to Unesco headquarters in Paris to push his case and plans to go to St. Petersburg. “For three months I couldn’t sleep,” Mr. Bader said. “I cannot imagine my village divided. If we have lived in peace these last 60 years, we can live the same way forever.”

Lung cancer “breathalyzer” wins $100K Entrepreneurship

  • MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    Photo: Michael Last


  • MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    Photo: Michael Last


  • MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    MIT $100K grand-prize-winning team Astraeus Technologies (left to right): Graham Lieberman, Jay Kumar, Alexander Blair, and Joseph Azzarelli.

    Photo: Michael Last


Lung cancer “breathalyzer” wins $100K Entrepreneurship Competition

Team’s smartphone-connected device can detect lung cancer early from a single breath.

Rob Matheson | MIT News Office 
May 12, 2016

A team of MIT and Harvard University students who invented a smartphone-connected sensor that detects lung cancer from a single breath took home the grand prize from Wednesday night’s $100K Entrepreneurship Competition.

Astraeus Technologies won the $100,000 Robert P. Goldberg Grand prize at the 27th annual competition, beating out seven other finalist teams that pitched business ideas to a panel of expert judges and a lively capacity crowd in Kresge Auditorium. Five other teams innovating in big data, creative arts, and food service took home separate category prizes totaling $40,000.     

Astraeus has developed a postage-stamp-sized device, called the L CARD, that detects certain gases indicative of lung cancer. When someone blows onto the device, a connected mobile app turns a smartphone screen red if those gases are present and green if they aren’t. 

Inventor Joseph Azzarelli, an MIT PhD student in chemistry, demonstrated the device on stage by spraying a syringe filled with the lung-cancer-signaling gases onto the device, causing the smartphone screen to flash red. “The L CARD reacts and sends instantaneous information to the physician that further attention is required,” Azzarelli said while a ripple of excitement spread through the crowd.

“We love that demo as much as you guys do,” added team member Jay Kumar, a student at Harvard Medical School.

After the competition, Azzarelli told MIT News the prize money will go toward product development and first-round clinical trials in research hospitals in the area.  

Cheaper, safer screening

Lung cancer is the deadliest type of cancer in the United States, causing more deaths than breast, colon, and prostate cancers combined, according to the World Health Organization. “Part of the reason lung cancer is so deadly is that the current gold standard screening test — the low-dose CT scan — is wholly inadequate in a variety of ways,” said team member Graham Lieberman, an MBA student at the Harvard Business School.

Kumar delved into more detail, explaining that CT scans cost about $800 for each scan, have a high false-positive rate, and expose patients to radiation that can increase their cancer risk.

Due to the risks and costs of CT scans, Lieberman added, only about 1.6 million of the 94 million Americans at risk for lung cancer — as estimated by the Centers for Disease Control and Prevention — are scanned each year. “A cheaper, safer screening device can be applied to a much larger percentage of that population,” he said.

The L CARD (which stands for Chemically Actuated Resonate Device) is essentially a modified near-field communication tag. Certain volatile organic compounds unique to the breath of lung cancer patients modify the tag’s radio frequency identification signal. A smartphone then pings the device and determines, from the modified signal, if those volatile compounds are present.

Kumar said the devices are an order of magnitude (about 10 times) more accurate than CT scans and can be made for less than $1. Astraeus will sell L CARDS directly to hospitals and clinics for use during routine annual checkups, he said. “We’re going after lung cancer,” Kumar said. “The root cause is bad screening: We’ve developed a better screening test, and it’s cost effective.”  

Last night’s win was the second for Astraeus in the $100K Entrepreneurship Competition, which consists of three independent contests: Pitch, Accelerate, and the Launch grand finale. Astraeus, which formed last November, also won the $10,000 Danny Lewin Grand Prize and the $3,000 Audience Choice Prize at the Accelerate competition on Feb. 10.

Going through the competition helped the team focus on all the steps it takes to establish and expand a business, Azzarelli told MIT News. “Going into the $100K — the Launch competition in particular, which is taken so seriously by so many — really forces you to think, ‘If we’re really going to do this, at the level we really want to do it at, how are we going to move forward,’” Azzarelli said.

Big winners

Several additional awards were granted last night to finalist and semifinalist teams: Finalist team Spyce, which developed a tumbler-type machine stocked with raw ingredients that autonomously cooks and serves meals in bowls to customers, won the $5,000 Audience Choice award.

A $10,000 data prize from Booz Allen Hamilton went to semifinalist team ReviveMed, which developed a platform that can be used to repurpose safe but shelved drugs at pharmaceutical firms, for other uses. Two teams split a $10,000 Thomson Reuters Data Prize: finalist teamHive Maritime, which is developing analytics and optimization algorithms for shipping routes and vessel speeds, based on predicted queues at ports and canals; and semifinalist teamSwift Calcs, which is creating a cloud-based computational platform for engineers to collaborate on calculations.

Taking home the $15,000 Creative Arts Prize was Tekuma, which developed a service that matches people who want to rent property with artists who create and curate art, and ships the art to the rented space.   

Five other finalist teams pitched ideas: AquaFresco developed a water-recycling technology that lets people use one batch of soapy water to clean their laundry for several months;DoneGood is an app that lets people rate businesses based on practices such as being green, supporting diversity, buying locally, and adequately supporting workers, among other causes; Lux Labs created a nanoscale film that selectively filters light to reduce energy consumption on mobile devices and improve efficiency of solar cells; Solugen invented a green, safe, scalable process for producing hydrogen peroxide, which is used for things like semiconductor fabrication, plastic production, and water purification; and ABA Power is making aluminum-based batteries that have 30 times the energy density of traditional lithium batteries and are manufactured with zero emissions.

Since its debut in 1990, the MIT $100K Entrepreneurship Competition has helped launch 160 companies worldwide that have raised an additional $1.3 billion in funding, have a combined market value of $16 billion, and have employed more than 4,600 people.

This year, 160 teams applied to the entrepreneurship competition. That number was winnowed to 50 semifinalist teams for the Launch contest. Judges then chose eight finalists to compete in Wednesday’s grand finale event. Semifinalist teams receive mentoring, prototyping funds, media exposure, and discounted services.

Tesla’s inherent safety saves five joyriding teenagers in Germany

Tesla’s inherent safety saves five joyriding teenagers in Germany

The car left the road, flew through the air and rolled into a field.

What’s left of the Model S after a teenager crashed and then rolled it into a field.
Sabine Hermsdorf

Do you have a teenage child that likes to borrow your car and then destroy it in a spectacular crash? We sincerely hope the answer to that question is a resounding “no,” but in the off chance that you do, you may want to consider changing your current vehicle for a Tesla Model S. Last week in Germany, the joyriding daughter of a Tesla owner discovered first-hand just how safe the electric vehicle is, after losing control at high speed and rolling into a field.

According to German newspaper Merkur, the 18-year old and four of her friends were messing around in her father’s Model S before losing control. The car flew more than 80 feet (25m) into a field before rolling once and coming to a halt. Although three of the occupants had to be helicoptered to hospitals in Munich for treatment, none of their injuries were life-threatening, a testament to the safety of Tesla’s skateboard chassis.

Enlarge / The skateboard chassis used by the Model S and Model X is extremely safe, with crumple zones that are unconcerned with engines that can transfer kinetic energy into the passengers during a frontal collision.
Jonathan Gitlin

Unlike a conventionally powered car, the Model S (and Model X) have no large engine up front to intrude into the passenger compartment during a collision. This means the front and rear crumple zones can effectively dissipate the kinetic energy of a crash, as seen to good effect in the photographs taken after the accident.

While the Model S was heavily damaged, one doesn’t need much of an imagination to think that a similar crash in a front-engined internal combustion vehicle would have had a much worse outcome for the car’s five occupants. (According to comments left at, about the only way to fatally crash a Tesla appears to be driving one off a cliff at high speed.)

Well done, Tesla.