BART, blockchain

Introducing BART, the Blockchain’s French Scientific Alliance

/0 Comments/in , /by

Four research institutions, including Télécom ParisTech and Télécom SudParis, are embarking on a joint research initiative focused on the blockchain. Along with INRIA and the Institute for Technological Research (IRT) SystemX, this scientific task force will take on the challenge of integrating the blockchain into industrial processes. The six focus areas of this research initiative called Blockchain Advanced Research & Technologies (BART) are scaling-up, security, data confidentiality, architecture, monitoring and business models. Gérard Memmi, head of the IT & Networks Department at Télécom ParisTech explains the objectives of this initiative, launched on 6 March 2018.

 

Why create a joint research initiative on the blockchain?

Gérard Memmi: The blockchain is often seen as a mature technology. In some ways it is, since it is based on scientific findings that have been known for 20 or 30 years—such as Merkle trees and the Byzantine generals problem. Yet this is not enough to make the blockchain the revolutionary technology that everyone expects it to become. Constant research must be carried out behind the scenes to ensure greater security, to scale the technology up for use by companies, and to make it compatible with current and future information systems. For example, the current proof-of-work system required for maintaining the blocks is incompatible with industrial performance requirements: it requires too many IT resources and too much time and energy. There are real scientific challenges ahead!

What will each of the different partners of the Blockchain Advanced Research & Technologies (BART) contribute?

GM: We expect INRIA to provide research in the areas of formal verification, theoretical computer science, cryptography… What I believe they expect of us is related to applications: connections with telecommunications protocols, the development of IT architectures, cybersecurity, scaling-up… However, I hope there will be a lot of interaction between the researchers of the different disciplines I just mentioned.  Through the involvement of IMT, we will also be able to use the TeraLab platform. SystemX contributes valuable connections with manufacturers and plans to use scientific findings to fuel work situated downstream in companies’ industrialization processes. That is their role as an Institute for Technological Research.

What is the goal of the Blockchain Advanced Research & Technologies (BART) initiative?

GM: In creating BART, we hope to form a multidisciplinary research team that will be a blockchain benchmark in France. We have the means to accomplish this partly because we have been working on this topic for a long time. Three years ago, we launched one of the first theses on the blockchain in France as part of the joint SEIDO laboratory with EDF, at a time when no one was talking about this technology. SystemX launched research projects involving industrial partners as early as 2016, specifically through the Blockchain for Smart Transactions (BST) project. We have therefore been able to begin creating a network of high-level scientific and industrial relationships.

Does this mean BART is anchored in dynamics that already existed?

GM: Yes, we are in the active launch phase. The signing of the framework agreement is not simply a matter of policy, it reflects real concrete actions. Personally, I see the BART initiative as part of a unified whole. For example, one of the PhD students funded by BART will go to Munich Technical University (TUM) to work on their blockchain platform. This cooperation really makes sense, since IMT and TUM have founded the Franco-German Industry of the Future Academy. In addition, as part of this alliance we are developing a Franco-German project called HyBlockArch devoted to blockchain architecture for industry. Although institutional connections do not always exist between these different entities, the work of each one is recognized by the others and contributes to the whole. This is also a great principle of international research: we come together with a common purpose and take advantage of all the possible synergies to make the greatest possible impact on science and society.

Also read on I’MTech: Is blockchain the ultimate technology of trust?

carnot TSN

Researching technological disruptions to prepare for the future

/0 Comments/in , /by

Belles histoires, Bouton, CarnotFundamental and applied research are often simplistically portrayed as being opposite each other. The Carnot program, run by the Ministry of Higher Education, Research and Innovation, and by the National Research Agency (ANR), challenges this vision. Though its primary objective is to develop virtuous partnerships between public research institutions and companies in order to stimulate technology transfer, it achieves this goal by requiring the institutions it funds to make significant contributions to fundamental research. Far from being paradoxical, this forward-looking strategy focuses on preparing companies for the future, beyond their immediate technological competitiveness. The Télécom & Société Numérique Carnot Institute (Carnot TSN, of which IMT is a member) fulfills this mission through its Futur & Ruptures (Future and Disruptions) program. Its director, Christian Picory-Donné, answers our questions to explain the scientific challenges of this type of program.

 

The Carnot program invites Carnot Institutes to carry out “scientific resourcing.” Can you explain what that is?

Christian Picory-Donné: It means preliminary research with the aim of preparing for the future of industry. What we have learned from the Fraunhofer Institutes in Germany, or other major technological research institutes, is that it is difficult for these organizations to be trailblazers. Their resources are targeted at satisfying nowadays market demands to such an extent that they tend to fail to prepare for the future. The Carnot program strives to address this shortcoming by providing funding well ahead of current industrial problems.

How does the Télécom & Société Numérique Carnot Institute (TSN) achieve this resourcing?

CP: The Futur & Ruptures program serves to fund the equivalent of 60 to 80 years of PhD, post-doctorate theses and sabbaticals every year. Funding for this support comes in part from Fondation Mines-Télécom, and in part from TSN Carnot. For TSN Carnot this initiative represents virtually all of its resourcing activity, or approximately 60% of the annual contribution it receives from the Carnot program. To give some concrete figures, this year TSN Carnot received a contribution of €4 million. This means €2.4 million was allocated to the Futur & Ruptures program to fund PhD and post-doctorate theses at members of TSN Carnot.

carnot TSN

Christian Picory-Donné, Director of TSN Carnot

The Carnot program’s objective is to develop partnership-based research. How does this upstream scientific positioning influence innovation?

CP: One example that comes to mind is the joint SePeMed laboratory between IMT Atlantique and the MEDECOM company, which was launched in 2014. It focuses on problems related to managing and securing medical databases. It all started with work by Gouenou Coatrieux, a researcher at IMT Atlantique, who obtained Carnot funding for his PhD and post-doctoral students through the Futur & Ruptures program. Thanks to the results achieved, he submitted a proposal for a Labcom (joint laboratory) project to the National Research Agency whose funding enabled a strategic partnership with MEDECOM through the creation of this joint laboratory was. There are many other success stories such as this one, and more to come since a great number of theses are funded while the benefits are not always as rapidly materialized or as directly related.

The Futur & Ruptures program is ten years old this year. What is your view of the program? 

CP: A very positive view. First of all, the Carnot program is a virtuous program, since it rewards research conducted by public laboratories in proportion to its effectiveness and the extent to which it fulfills its commitments to expanding partnership-based research. It is also a tool to support IMT’s strategy. Originally focused on research to support economic development, the Carnot program also supports IMT’s development plan. The same is true for the tools supported by the Foundation, so much so that a significant leverage effect can be observed—funding, development of research resources, research results—in keeping with the institute’s strategic priorities.

This reality of partnership-based research contrasts with the stereotype of research activity as a service provided for a company.   

CP: Of course. We consider it to be targeted research since it is attentive to companies’ problems, but that does not mean there is a subordinate relationship. Resourcing may be seen from different perspectives. First of all, there is this noble image of the researcher toiling away in his ivory tower to achieve his vision. This is not our standpoint, because Carnot resourcing is more connected to the vision we may have about long-term industrial needs. It is a very forward-looking position. We know, for example, that data protection and cryptography are a major issue—simply by being attentive to current industry concerns—so we have researchers working on quantum cryptography. But no company has come to ask us to fund such specialized research for use in immediate applications. That would be too risky for companies: industrial needs for quantum cryptography will not become a reality until a number of years from now. This positioning allows us to carry out fundamental research, while looking ahead to the technological obstacles that companies will face in the future.

[box type=”shadow” align=”” class=”” width=””]

The TSN Carnot institute, a guarantee of excellence in partnership-based research since 2006

Having first received the Carnot label in 2006, the Télécom & Société numérique Carnot institute is the first national “Information and Communication Science and Technology” Carnot institute. Home to over 2,000 researchers, it is focused on the technical, economic and social implications of the digital transition. In 2016, the Carnot label was renewed for the second consecutive time, demonstrating the quality of the innovations produced through the collaborations between researchers and companies.

The institute encompasses Télécom ParisTech, IMT Atlantique, Télécom SudParis, Télécom École de Management, Eurecom, Télécom Physique Strasbourg and Télécom Saint-Étienne, École Polytechnique (Lix and CMAP laboratories), Strate École de Design and Femto Engineering. Learn more [/box]

Photomécanique Jean-José Orteu

What is photomechanics?

/0 Comments/in , , , /by

How can we measure the deformation of a stratospheric balloon composed only of an envelope a few micrometers thick? It is impossible to attach a sensor to it because this would distort the envelope’s behavior… Photomechanics, which refers to measurement methods using images and computer analysis, makes it possible to measure this deformation or a material’s temperature without making any contact. Jean-José Orteu, a researcher in artificial vision for photomechanics, control and monitoring at IMT Mines Albi, explains the principles behind photomechanical methods, which are used in the aeronautics, automotive and nuclear industries.

 

What is photomechanics?

Jean-José Orteu: We can define photomechanics as the application of optical measurements to experimental mechanics and, more specifically, the study of the behavior of materials and structures. The techniques that have been developed are used to measure materials’ deformation or temperature.

Photomechanics is a relatively young discipline, roughly 30 years old. It is based on around ten different measurement techniques that can be applied to both a nanoscale and the dimensions of an airplane and to both static and dynamic systems. Among these different techniques, two are primarily used: the digital image correlation (DIC) method for measuring deformations, and the infrared thermography method for measuring temperatures.

 

How are these two techniques implemented?

JJO: For the DIC, we position one or several cameras in front of a material: only one for a planar material that undergoes in-plane deformation and several for the measurement of a three-dimensional material. The cameras film the material as it is deformed under the effect of mechanical stress and/or heat. Once the images are taken, the deformation of the material is calculated based on the deformation of the images obtained: if the material is deformed, so is the image.  This deformation is measured using computer processing and is extrapolated to the material.

This is referred to as the white light method because the material is lit by an incoherent light from standard lighting. Other more complex photomechanical techniques require the use of a laser to light the material: these are referred to as interferometric methods.  They are useful for very fine measurements of displacements in the micrometer or nanometer range.

The second most frequently used technique in photomechanics is infrared thermography, which is used to measure temperatures. This uses the same process as the DIC technique, with the initial acquisition of infrared images followed by the computer processing of these images to determine the temperature of the observed material. Calculating a temperature using an image is no easy task. The material’s thermo-optical properties must be taken into account as well as the measuring environment.

With all of these techniques, we can analyze the dynamic evolution of the distortion or temperature. The material is therefore analyzed both spatially and temporally.

photomécanique Jean-José Orteu

photomechanics, Jean-José Orteu

Stereo-DIC measurement of the deformation field of a sheet of metal shaped using incremental forming

 

What type of camera is used for these measurement methods?

JJO: While camera resolution influences the quality and precision of the measurements, a traditional camera can already obtain good results. However, to study very fast phenomena, such as the impact of a bird in flight on an aircraft fuselage, very fast cameras are needed, which can take 10,000, 100,000 or even 1,000,000 images per second! In addition, for temperature measurements, infrared-sensitive cameras must be used.

 

What is the value of optical measurements as compared to other measurement methods?

JJO: Traditionally, a strain gauge is used to measure the deformation of a material. A strain gauge is a sensor that is glued or welded to the surface of the material to provide an isolated indication of its deformation. This gauge must be as nonintrusive as possible and must not alter the object’s behavior. The same problem exists for temperature measurements. Traditional techniques use a thermocouple, a temperature sensor that is also welded to the surface of the material. When the sensors are very small compared to the material, they are nonintrusive and therefore do not pose a problem. Yet for some applications, the use of contact sensors is impossible. For example, at IMT Mines Albi we worked on the deformation of a parachute when it inflates. But the canvas contained a lining only a few micrometers thick. A gauge would have been difficult to glue to it and would have greatly disrupted the material’s behavior. In this type of situation, photomechanics is indispensable, since no contact is required with the object.

Finally, both the gauge and the thermocouple offer only isolated information, only at the spot where the sensor is glued. You won’t get any information concerning a spot only ten centimeters away from the sensor. However, the problem in mechanics is that, most of the time, we do not know exactly where we will need information about deformation or the temperature. The risk is therefore that of not welding or gluing the sensors in the spots where the deformation or temperature measurement is the most relevant. The optical methods also offer field information: a deformation field or temperature field.  We can therefore view the material’s entire surface, including the areas where the deformation or temperature gradient is more significant.

 

Photomécanique Jean-José Orteu

Photomécanique Jean-José Orteu

phtomechanics

Top, a material instrumented with gauges (only 6 measurement points). Middle, the same material to which speckled paint has been added to implement the optical DIC technique. Bottom, the deformation field measured via DIC (hundreds of measurement points).

What are the limitations of photomechanics?

JJO: In the beginning, photomechanical methods based on the use of cameras could not measure surface deformations. But over the last five or six years, an entire segment of photomechanics has begun to focus on deformations within objects.  These new techniques require the use of specific sensors, tomographs. They make it possible to take X-ray images of the materials, which reveal core deformations after computer processing. The large volumes of data this technique generates raise big data issues.

In terms of temperature, the core measurement without contact is more complicated. We recently defended a thesis at IMT Mines Albi on a method that makes it possible to measure the temperature in a material’s core based on the fluorescence phenomenon. The results are very promising, but the research must be continued to obtain industrial applications.

In addition, despite its many advantages, photomechanics has not yet fully replaced strain gauges and thermocouples. In fact, optical measurement techniques have not yet been standardized. Typically, when measuring a deformation with a gauge, the method of measurement is standardized: what type of gauge is it?  How should it be attached to the material? A precise methodology must be followed. In photomechanics, whether in the choice of camera and its calibration and position, or the image processing in the second phase, everything is variable, and everyone creates his or her own method. In terms of certification, some industrial stakeholders therefore remain hesitant about the use of these methods.

There is also still work to be done in assessing measurement uncertainties. The image acquisition chain and processing procedure can be complex, and errors can distort the measurements in any stage. How can we ensure there are as few errors as possible? How can we assess measurement uncertainties? Research in this area is underway. The long-term goal is to be able to systematically provide a measurement field with a range of associated uncertainties. Today, this assessment remains complicated, especially for non-experts.

Nevertheless, despite these difficulties, the major industries that need to define the behavior of materials, such as the automotive, aeronautics and nuclear industries, all use photomechanics. And although progress must be made in assessing measurement uncertainties and establishing standardization, the results these optical methods achieve are often of better quality than those of traditional methods.

 

TeraLab, a European data sanctuary

/0 Comments/in , , /by

Projets européens H2020Over the course of three months, TeraLab was involved in two European H2020 projects on the industry of the future: MIDIH and BOOST 4.0. This confirms the role played by TeraLab—IMT’s big data and artificial intelligence platform—as a trusted third party and facilitator of experimentation. TeraLab created a safe place for these projects, far from competitive markets, where industry stakeholders could accept to share their data.

 

“Data sharing is the key to opening up research in Europe,” says Anne-Sophie Taillandier. According to the director of TeraLab—IMT’s big data and AI platform—a major challenge exists in the sharing of data between industrial players and academics. For SMEs and research institutions, having access to industrial data means working on real economic and professional problems.  This is an excellent opportunity for accelerating prototypes and proofs of concept and removing scientific barriers. Yet for industrial stakeholders, the owners of the data, this sharing must not compromise security. “They want guarantees,” says the Director, who was ranked last February among the top 20 individuals driving AI in France by French business magazine L’Usine Nouvelle.

It is in this perspective of offering guarantees that the TeraLab platform joined the consortia of two European projects from the H2020 program: BOOST 4.0 (January 2018) and MIDIH (October 2017). The first project brought together 50 industrial and academic partners, including 13 pilot plants in Europe. The project is intended to create a replicable model of a smart industry in which data would form the basis for reflections on operational efficiency, user experience and even the creation of a business model. This level of ambition requires significant work on interoperability, security and data sharing. “But it is clear that Volvo and Volkswagen, both members of the Boost 4.0 consortium, will not provide access to their data without first experiencing a certain level of trust,” explains Anne-Sophie Taillandier. A platform like TeraLab allows companies to benefit from technological and legal advantages that make it a safe workspace.

The MIDIH project, on the other hand, seeks to provide companies with the technological, financial and material resources required for developing innovative solutions for the industry of the future through sub-grants. “In practical terms, the H2020 project will finance calls for projects on logistics, predictive maintenance and steel cutting and will offer support to successful applicants,” the TeraLab director explains. The companies selected through these calls for projects will be able to develop proofs of concept for solving industrial problems experienced by SMEs. They use platforms like TeraLab to accomplish this, since they “provide the assurance of the sovereignty and cybersecurity of the data the prototypes will produce.” For these companies, the ability to use an independent platform of this magnitude is truly beneficial in accelerating their projects.

A platform recognized at European level

TeraLab’s involvement in these projects is also due to the recognition it has earned at European level. In 2016, the Big Data Value Association (BDVA) granted TeraLab its Silver i-Space Label. This recognition is far from trivial, since BVDA leads the European private-public partnership on big data. BOOST 4.0 is the result of reflection carried out by this same partnership, which works to advance the major issues that industrial stakeholders have presented to the European Commission. “The context of the European Commission is incredible because many different stakeholders gravitate there, but within a given theme, everyone knows each other,” Anne-Sophie Taillandier admits. “The Silver i-Space Label awarded in 2016 provided both recognition from big data stakeholders and strategic positioning within this environment.”

In Europe, few platforms like TeraLab exist. Only ten hold the Silver i-Space Label—the highest level—held by TeraLab, the only French awardee of this recognition. It therefore represents a valuable gateway to involvement in European projects. “It legitimizes our responses to calls for bids such as these two projects on industry 4.0,” says the Director of the platform. The industry of the future is a topic TeraLab had already worked on before joining the MIDIH and BOOST 4.0 projects. “One of our strengths, which was recognized by both consortia, was our ability to develop a community of researchers and innovators on this subject,” says Anne-Sophie Taillandier. She also reminds us that the industry is not the only theme TeraLab has explored in the context of in-depth projects. This offers good prospects for TeraLab to be involved in other European projects on other specialized areas, such as healthcare.

 

AiZimov

Adrock.Tv, AiZimov and Seaclick: three new startups get interest-free loans

/0 Comments/in , /by

On April 5th, the approval committee for the Digital Fund of the Graduate Schools and Universities Initiative chose three new startups from IMT incubators to receive interest-free loans: Adrock.tv and AiZimov, from the ParisTech Entrepreneurs incubator and Seaclick, from the IMT Starter incubator.

These loans aimed at financing the development of young promising companies have a 0% interest rate and can be awarded for amounts up to €60,000. They are co-financed by Fondation Mines-Télécom, the Caisse des Dépôts and Revital’Emploi.

 

[one_half][box type=”shadow” align=”” class=”” width=””]start up adrock.tvAdrock.tv offers an artificial intelligence tool that analyzes images from editorial content online and integrates relevant ads from advertisers. [/box]

[box type=”shadow” align=”” class=”” width=””]start up seaclick

The online tool Seaclick makes it possible to view and buy tickets to local cultural and sports events in just a few clicks.[/box]

[/one_half]

[one_half_last][box type=”shadow” align=”” class=”” width=””]Aizimov Start up

AiZimov develops artificial intelligence for sales representatives that lets them select pertinent profiles online and write personalized emails based on the individual’s circumstances.[/box][/one_half_last]

 

startups, start-up, prêts d'honneur, Fondation Mines-Télécom, interest-free loans

12 startups supported with interest-free loans in 2017

/0 Comments/in , , /by

Thanks to its generous sponsors, Fondation Mines-Télécom was able to fund 22 interest-free loans in 2017 for entrepreneurs supported by IMT incubators. A total of €440,000 was awarded to 12 startups. In 2018, the Foundation intends to take its support for entrepreneurship a step further by awarding 30 interest-free loans.

 

Promising figures

To support the development of startups at IMT incubators, Fondation Mines-Télécom awards interest-free loans to entrepreneurs through the Graduate Schools and Universities Initiative. In 2017, twelve startups selected by the Foundation’s corporate partners benefitted from an interest-free loan. A total of €2.3 million in loans has been awarded since 2008, with an attrition rate of under 8%.

These no-collateral loans ranging between €20,000 and €40,000 help leverage funding for projects. Startups that received these loans have raised considerable funds this year, especially fintech Pledg (€1.2 million) and Seaver, which specializes in the IoT (Internet of Things) for the equine industry. The fact that many of these startups take part in the Las Vegas CES every year also attests to their strong performance. The objective for 2018 is to award interest-free loans to 30 new entrepreneurs to support 15 projects, representing a total of more than €560,000. To help achieve this goal, program partners Caisse des Dépôts and Revital’emploi are renewing their support.

Innovative services and products in the field of digital technology

The startups supported by these loans respond to different needs in the field of digital technology and take advantage of opportunities provided by big data. One such startup, Predictice, provides analytic solutions for court decisions designed for legal professionals.

Many of the startups specialize in connected objects. HEROZ, for example, is a connected accessory that protects smartphones from being stolen, lost, forgotten and also protects against intrusion. Keepen offers an autonomous alarm that everyone can access which is both more convenient and more reliable than current systems.

They also provide innovative services and create new user experiences. ThingType provides an online service for electronic design and creation that makes prototyping simple, easy and affordable, while Bruce, a digital and mobile temporary employment agency, helps businesses meet their needs for temporary employment.

Find out more about the interest-free loan program

Nearly 100 startups and spinoffs are created through the IMT incubator network every year. This is why Fondation Mines-Télécom finances a portion of the Digital Fund of the Graduate Schools & Universities Initiative, part of the Initiative France network, which aims to foster the development of new businesses at graduate schools, universities and research laboratories in France.

The interest-free loans play a key role in the development of startups that strive to expand rapidly, both in France and internationally. The loans provide the project with legitimacy, are accompanied by the incubators’ technical expertise and help leverage funding for startup costs. These loans for an amount of up to €40,000 have a one-year grace period and must then be repaid within five years. Repayments are in turn used to provide funding for loans for other entrepreneurs.

[divider style=”dotted” top=”10″ bottom=”10″]

The original version of this article was published on the  Fondation Mines-Télécom website

[divider style=”dotted” top=”10″ bottom=”10″]

facial biometrics

Facial biometrics: How smartphones can recognize us

/1 Comment/in , /by

Mohamed Daoudi, IMT Lille Douai – Institut Mines-Télécom

[divider style=”normal” top=”20″ bottom=”20″]

[dropcap]W[/dropcap]elcome to the new era: that of facial biometrics. The launch of the iPhone X, a smartphone featuring Face ID facial recognition, demonstrated that this technology has now reached full maturity. This became possible with the introduction of miniature 3D sensors with high-level computing power, combined with extremely efficient learning algorithms such as deep learning.

But what is facial recognition? It means identifying that two faces are identical despite changes caused by lighting conditions, pose and facial expressions. Generally speaking, this means finding distances within the face that can be used to identify any changes to the face.

Figure showing the same face in different shooting conditions and lighting changes.

 

In 2014, researchers from Facebook published an article called “DeepFace: Closing the Gap to Human-Level Performance in Face Verification”. To prevent the problems caused by changes in pose, a step was introduced to align the 2D face to a 3D model of the face. The next step involved a deep learning process using a network of artificial neurons consisting of 120 million connections. The learning set was composed of 4.4 million faces of celebrities. The network of neurons was trained to recognize the variances in the faces. The algorithm made it possible to determine if two photographed faces belonged to the same person with a specified accuracy of 97.35%.

In 2015, researchers from Google published an article entitled “FaceNet: A Unified Embedding for Face Recognition and Clustering”. They showed that they were able to achieve a recognition rate of 99.63% using a database of 2D faces captured in an uncontrolled environment. To accomplish this, the authors proposed the use of a neural network consisting of eleven convolutional layers and three connected layers. The idea was to ensure that an image of a specific person would be closer to all the other images of that same person (referred to as positive) than to the images of other people (referred to as negative). The learning was carried out using a database of 200 million face images from 8 million people.

facial biometrics

During the training, the learned similarities allowed the images showing the same faces to come closer together, and those showing different faces moved farther apart in relation to a specific metric.

 

However, the DeepFace and FaceNet experiments were both based on private databases that are not available to the scientific community. A team from the University of Oxford proposed to collect data from the web and has established a database of 2.6 million faces from 2,622 people and has proposed a network architecture called VGG-face consisting of 16 convolutional layers and 3 fully connected layers. Today this architecture is widely used by the computer vision community.

Yet the face is not only a 2D image; it is also a three-dimensional image. Facial biometrics can be used because 3D scanning technologies can scan faces. The major advantage of using 3D in this context is that the facial recognition algorithms are resistant to changes in lighting and pose. Recent work published in 2013 by our team at IMT Lille Douai in the journal IEEE TPAMI, “3D face recognition under expressions, occlusions, and pose variations” showed the advantage of this process. In this article, we proposed to compare two 3D faces by comparing two sets of curves that locally represent the shape of a 3D face. We obtained a recognition rate of 97% (using the testing framework Face Recognition Grand Challenge). The results obtained from several international tests reveal the advantages of 3D faces in facial biometrics systems.

Example of 3D faces captured by the Minolta scanner using laser technology.

 

Now let us get back to the iPhone X and its 3D technology for facial recognition. A feat made possible by the introduction of miniature 3D sensors on the front of the device: a projector sends 30,000 invisible points onto the user’s face, which are used to create a 3D model of the face. According to Apple, Face ID cannot be fooled by a mere photograph of a face, since the recognition is achieved with a 3D sensor that measures depth.

Mohamed Daoudi, Professor at the IMT Lille Douai, Lille Center of Research in Computer Science, Signal and Automatic Control, IMT Lille Douai – Mines-Télécom Institute

The original version of this article was published in French on The Conversation.

watermarking

Watermarking: a step closer to secure health data

/0 Comments/in , , , /by

Belles histoires, Bouton, CarnotIn the near future, watermarking data could be the best traceability technique in the healthcare domain. It involves hidding information into medical images with the aim at reinforcing data security for patients and healthcare professionals. After being developed for nearly ten years in the laboratories of IMT Atlantique and Medecom, watermarking has now reached a level of maturity that allows its integration into professional products. Yet it still must be approved by standardization bodies.

 

Are you sure that is really your body on the latest X-ray from your medical exam? The question may seem absurd, yet it is crucial that you, your doctor and the radiologist can all answer this question with a resounding “yes”. To ensure this level of certainty, healthcare professionals must rely on the latest technological advances. This is a matter of ensuring the right patient gets the right diagnosis—no one wants an X-ray of their lungs to be switched with one from a chain-smoker!

To ensure an X-ray is correctly associated with the patient and to return a lost X-ray to its rightful owner, the name printed on the X-ray film is not sufficient. An ill-intentioned individual or an administrative error could cause the unfortunate exchange of two patients’ images. Medecom and researchers from IMT Atlantique, part of the Télécom & Société Numérique Carnot Institute, have been working on a more secure system based on the watermarking. For over ten years the two entities have been collaborating on this technology and four years ago they inaugurated SePEMeD, a joint laboratory focused on this area, with support from the French National Agency for Research (ANR).  Since then, the maturity and viability of the watermark technology have become increasingly convincing.

A Secret Message

“The watermark draws on the principle of steganography, the art of hidden writing, which is almost as old as cryptography,” explains Gouenou Coatrieux, a researcher in imagery and information processing at IMT Atlantique. “In the case of X-rays, we change some pixels in the image to hide a message and leave an invisible mark,” he continues (see box below). The value of the watermark is that the protection is independent from the storage format. The X-ray can therefore be exchanged between departments and hospitals, each with its own unique system for processing X-ray images, yet this will not affect the watermark, which will continue to contain the information related to the patient.

[box type=”info” align=”” class=”” width=””]

The watermark: a message hidden in the pixels

The secret to watermarking X-rays is in the pixels, which can be encoded in 8, 16 or 32 bits. If a pixel is encoded in 8 bits, this means its color is indicated by a series of 8 bits—a bit is a 0 or 1 in the binary code. There are 256 possible 8-bit combinations: 00000000, 00000001, etc. There are therefore 256 possible colors for a pixel encoded in 8 bits, or 256 different shades of gray for a pixel in a black and white image.

Watermarking an image involves modifying certain pixels by changing one of their bits. This means the color, or shade of gray, is altered. To prevent this from being noticeable on the X-ray, the bit containing the least amount of information—the one located at the end of the 8-bit sequence—is modified. The colors related to bits 00110101 and 00110100 are very similar, whereas those related to bits 00000000 and 10110110 are very different. The more two series of bits are dissimilar, the less similar the colors.

The changed bits in the pixels form a message, which could be a patient’s name or the doctors’ authorization to access the X-ray. To discover which bits bear this message, the X-ray recipient must have the watermark key associated with the medical image. This ensures the secrecy of the message.[/box]

In addition to traceability, watermarking has other advantages. First, it can help detect insurance fraud. If an X-ray is tampered with by an ill-intentioned individual, for example to fake a disease, the secretly watermarked pixels will also be modified, revealing the attempted fraud. Next, the watermark can be added to data that is already encrypted using a method that has been patented by Medecom and IMT Atlantique. It is therefore possible to ensure traceability while maintaining the confidentiality of medical information the image contains. This also makes it possible to write information about certain doctors’ access authorizations directly on the encrypted data.

Moving towards standardization?

While this watermark technology is now mature, it still must pass the test of standardization procedures in order to be implemented in software and the information systems of healthcare professionals. “Our goal now is to show that altering the image with the watermark does not have any effect on the quality of the image and the doctors’ diagnostic capacity,” says Michel Cozic, R&D director at Medecom. The SePEMeD team is therefore working to conduct qualitative studies on watermarked data with physicians.

At the same time, they must convince certain healthcare professionals of the value of watermarking.  The protection of personal data, and medical data in particular, is not always viewed the same way throughout the healthcare world. “In the hospital environment, professionals tend to believe that the environment is necessarily secure, which is not always the case,” Michel Cozic explains. In France, and in Europe in general, attitudes about data security are changing. The new General Data Protection Regulation (GDPR) established by the European Commission is proof of this. However, it will be some time before the entire medical community systematically takes data protection into account.

Ten years of research… and ten more to come?

Since there is still a long way to go before healthcare professionals begin using watermarks, the SePEMeD story is not over yet. Founded in 2014 to solidify the collaboration between IMT Atlantique and Medecom, which has lasted over ten years, SePEMeD was originally intended to run only three years. However, following the success of the research which led to promising applications, this first joint laboratory accredited by the ANR on data security will continue its work until at least 2020. Beyond data traceability, SePEMeD is also seeking to improve the security of remotely processed encrypted images in cloud storage.

“We update our focus areas based on our results,” Gouenou Coatrieux notes, explaining why the SePEMeD laboratory has been extended. Michel Cozic agrees: “We are currently focusing our research on issues related to browsers’ access to data, and the integration of watermarking modules in existing products used by professionals.” The compatibility of algorithms with healthcare institutions’ computer configurations and systems will be a major issue involved in the adoption of this technology. Last but not least: ease of use.  “No one wants to have to enter passwords in the software,” observes Medecom’s R&D Director. We must therefore succeed in integrating watermarking as a security solution that is straightforward for doctors.

[divider style=”normal” top=”20″ bottom=”20″]

The benefit of collaborating with IMT Atlantique: “The human aspect”

Michel Cozic

One of Télécom & Société Numérique Carnot Institute’s objectives is to professionalize the relationships between companies and researchers.  Michel Cozic, Medecom’s R&D Director shares his experience: “There is also a human aspect to these collaborations. Our exchanges with IMT Atlantique go very smoothly, we understand each other. On both sides we accept our differences, constraints and we compromise. We come from two different environments and this means we must have discussions. There must be an atmosphere of trust, a good relationship and a common understanding of the objectives. This is what we have been able to accomplish through the SePEMeD laboratory.”

[divider style=”normal” top=”20″ bottom=”20″]

[box type=”shadow” align=”” class=”” width=””]

The TSN Carnot institute, a guarantee of excellence in partnership-based research since 2006

Having first received the Carnot label in 2006, the Télécom & Société numérique Carnot institute is the first national “Information and Communication Science and Technology” Carnot institute. Home to over 2,000 researchers, it is focused on the technical, economic and social implications of the digital transition. In 2016, the Carnot label was renewed for the second consecutive time, demonstrating the quality of the innovations produced through the collaborations between researchers and companies.

The institute encompasses Télécom ParisTech, IMT Atlantique, Télécom SudParis, Télécom École de Management, Eurecom, Télécom Physique Strasbourg and Télécom Saint-Étienne, École Polytechnique (Lix and CMAP laboratories), Strate École de Design and Femto Engineering.

[/box]

Franco-German Academy, Industry of the Future

Industry of the Future: How is the German-French Academy supporting European companies?

/1 Comment/in /by

Opinion. By Hannemor Keidel, TUM Officer for Scientific Alliances with France, and Christian Roux, Executive Vice President for Research and Innovation at IMT. This article is the long version of the editorial published by Innovation Review. 

[divider style=”normal” top=”20″ bottom=”20″]

[dropcap]T[/dropcap]he German-French Academy for the Industry of the Future was founded a little over two years ago and the first actions it has supported are now taking shape. Research, innovation and training are the three components that guide the work of this transnational institution created by IMT and Technische Universität München (TUM). The ambitious strategy behind this emerging project is to support European companies in the major changes driven by the digital transition.

 

The explosion of digital technologies offers an advantage for companies’ competitiveness. The world’s major economic areas—including Europe—have understood this. In supporting the transformation of their industries, they see the opportunity to gain technological leadership and pursue their development, a goal that cannot be reached without the use of disruptive technology and new professional skills. In this area, academic institutions are key. They offer a decisive contribution in their ability to carry out forward-looking analysis and in the technological and educational innovations they develop.

As for many strategic issues in Europe, the cooperation between France and Germany serves as a launchpad for major structural projects. IMT and Technische Universität München (TUM) have sought to further encourage this emulation process by forming a partnership with a Franco-German academy devoted to the industry of the future. Its aim is to create a cutting-edge European institution based on a new model in research, innovation and training for the industry of the future.

The programs they develop will embody the potential of the economy and industry to transition and become digitized while truly considering companies’ real needs. In practical terms, this action will take the form of transnational industrial Chairs, new initial and lifelong training opportunities for professionals, networks of incubators and technology platforms.

Not merely a vision: concrete action

Five research projects related to the scientific component were launched in June 2017. The projects can be categorized into two themes. On the one hand, HyBlockArch and Industry without Borders are studying network cooperation for the industry of the future. On the other hand, the SCHEIF, SeCIF and ASSET projects are focusing on cyber-architecture and cyber-security. The exploration of scientific fields will expand to include other themes of the future: advanced materials, additive manufacturing, energy, industrial logistics, predictive maintenance…

In the area of training, the German-French Academy has also developed initiatives. Two summer schools were held in 2017: one at EURECOM in France on human factors and human-machine interaction and another in Munich on mobility and smart roads.  New summer schools for PhD students and research professors will also be offered on new themes. In addition, a wide range of training (initial, ongoing and lifelong learning) will be developed to offer solutions to companies in key sectors that are the most affected by digitization. This training will be supported by a selection of MOOCs devoted to the industry of the future.

Finally, the Academy’s third component, innovation, will be launched in the spring of 2018. IMT and TUM will offer companies joint services that will give them access to the very best in academic research and innovative systems. By allowing economic actors to access leading-edge technology platforms, IMT and TUM are seeking to foster a value within the Academy for supporting the economy and French and German manufacturers. This will be another step forward in this joint development project for the future of European industry.

 

IMT and TUM create the german-french academy for industry of the future