Introduction A brief background
p.3 p.4

Performance… and comparison with other forms of biometrics p.14 Implementing a facial recognition system

- The origins - Facial recognition with good quality portraits - General facial recognition

The applications
- Criminal Justice


-  Integrating facial recognition in the existing technical environment -   Managing the expectations of customers and operators -  Developing the expertise of operators
- Expertise in data acquisition - Expertise in checking searches

- Identification and maintenance of a portrait reference database - Identity checks in the field - Criminal investigations and information - Prevention

-  Helping the end users

-  Civil applications, access control and border controls
- Issue of identity cards - Control of identity cards - Access control

The market associated with automatic facial recognition 
The market for facial recognition -  Driving forces and obstacles


-  Applications for the general public
- Access to computerized services - Photo album management

The experience of Morpho in the field of facial recognition p.20

The Technology
-  The different steps of automatic facial recognition

Glossary and acronyms Bibliography

p.22 p.23

- Step 1 - image acquisition - Step 2 - face searches with scaling and alignment - Step 3 - enhancement of face images - Step 4 - extraction of characteristics - Step 5 - representation as a template and comparison - Step 6 - thresholding and decision-making

-  The facial recognition algorithms
- Procedural algorithms - Training algorithms 

Special features of video processing -  Special features of 3D - Light sources


This apparent ease of use is the cause for many fantasies. the human brain and visual organs can be considered the best existing face recognition machine – ever – and forever. Human face recognition is the most widely used way of identification or authentication of identity. Because face recognition is the most natural thing for any human being. in order to deliver its full efficiency. since the human face easily leaves “traces” when crimes are recorded by CCTV cameras or the cameras of witnesses. Rapid advances are being made in facial recognition technology. driver license. facial recognition would appear to be the most natural of biometric techniques. a number of caveats have to be taken into account during implementation. A specific area of the human brain called fusiform face area (FFA) has been proven to be totally dedicated to this task. library card or gym club card. up to the point where it is on the verge of taking on its greatest challenge: beat the human brain. trained from birth to somehow do the exact same thing There are many applications. its use. the state of the art of the technology and its technical and commercial potential. best face recognition machine ever. Because the technology has an outstanding competitor : the human brain. and accounts for the presence of portraits on most of the identification documents that we carry in our purses and wallets. Automatic facial recognition is actually used in civil spheres in order to guarantee the unique nature of identity documents and in military or law enforcement applications.Introduction For obvious reasons. Still. but not forever. 3 . In an effort to answer the increasing number of questions being put forward. be it an ID card. contactless and does not require any highly specific equipment. credit card. a fact that facilitates the implementation of automatic facial recognition. The acquisition of portrait images is simple. inconsistencies and difficulties while implementing automatic face recognition systems in the field. Morpho has taken stock of facial recognition in general. and it thus has every trait expected of a major biometric technique.

facial recognition was already the subject of a great deal of research.). the eyes.g. because faces are by nature very mobile and measurements between characteristic points are affected by orientation.e.A brief background The origins From the very advent of photography. Figure 1: Portrait parlé “class”. They essentially consisted in checking the coherence of measurements between different characteristic points of the face (e. Examples include: • the development of identification parade or “lineup”(1) techniques in the United Kingdom. In order to recognize delinquents who are repeatedly arrested.g. Researchers quickly found that the overall issue of facial recognition was complex. At the end of the 1980s. Facial recognition with good quality portraits The first attempts to automate facial recognition started in the 1960s in semi-automatic mode. the hairline. Before the use of computers to recognize faces was even considered a possibility. This technique is used to find a face in a photo and to compare images of faces. without having to resort to large collections of portraits. Research had reached a point where the operational use of facial recognition on high-resolution frontal images taken in a controlled environment was now feasible. 1853-1914. Its conclusions were quite clear. 4 . At the start of 2007. in which a witness is confronted with a group of physically similar people. the corners of the eyes. criminologist who developed judicial anthropometry in France. Source Library of Congress. They were not very successful. but could be simplified by only taking into consideration portraits that are coherent in terms of orientation. the ICAO* defined criteria to obtain controlled portraits and meaningful test sets were created. nose or mouth). but facial recognition has thus become a biometric technique in its own right. i. This classification is known as the “spoken portrait”. the specific shapes of the different parts of the face. Research focused on this problem. (1) Alphonse Bertillon. (2) Eigenfaces: a facial recognition technique that consists in learning the distinctive characteristics of faces from a broad sample of portraits using each complete image rather than local characteristics (e. But this event obviously did not put an end to work on the recognition of controlled portraits. to the extent that speciallydeveloped models quickly proved to be necessary. one of whom is a suspect. both government agencies and private organizations have kept collections of portraits and ID photos have gradually made their way onto all personal identification documents. etc. the NIST* published the results of its “FRVT 2006”* test. expression and image quality. • the work done by Bertillon on face classification. The witness must decide whether one of the persons in the group was present at the scene of the crime. lighting. More improvements are expected. from the most official passports to informal membership cards issued by sports clubs. the development of the eigenfaces(2) technique prompted a more intense research effort. USA. Bertillon suggested that the portraits be sorted by common morphological characteristics.

5 . research has been looking into significantly more difficult problems. in which faces are not viewed frontally. USA. It is the start of a new era and we can expect to see significant progress over the coming years. courtesy of Santiago Serrano Drexel University. Figure 3: Facial recognition history. resolution is low or the image quality is sometimes poor. Figure 2: Eigenfaces.General facial recognition Since 2007. the NIST is again seeking to assess performance and has provided researchers with representative data (images and videos of faces under non-controlled conditions). With the MBGC*.

Criminal Justice Identification and maintenance of a portrait reference database Just like automatic fingerprint recognition. Facial recognition is special due to the portraits themselves. Identity checks in the field With just a camera and suitable means of transmission.The Applications Automatic facial recognition is a form of biometrics. facial recognition provides more benefits: • it increases population coverage of the identification scheme. facial recognition allows police forces to manage the files of people of interest by making sure that there are not several different records for a single person. facial recognition has applications in the policing and civil fields and for access control. It is used for authentication (checking that a person really is who they say they are) and identification (finding out who someone is from a group of known persons). demonstrated the high precision of the automatic facial identification of suspects and that identification is possible without calling on fingerprint experts. While this task is already performed using fingerprints. superior identification performance can be achieved. Figure 4: mobile facial recognition. The combination of the two biometric techniques increases the precision of searches and allows reliable. Police officers equipped with PDAs can submit search requests to remote facial recognition systems and quickly determine whether an individual is already known to the forces of law and order. USA. 6 . automatic decisions to be sent to the field. which are widely available and easy to acquire. if the officer has the equipment required to take fingerprints. enabling identification of individuals whose fingerprints cannot be acquired for various reasons • by combining the two biometric modalities. without the officer requiring any expertise in fingerprints. therefore reducing the workload involved in the verification process The Pierce County Sheriff’s Office in Washington. Like most biometric techniques. Their use is acceptable to the general public. it is possible to check the identity of a person in the field using a photograph of their face. ID checks can be carried out on just the face or both the fingerprints and the face.

Prevention Facial recognition can also be used for preventive purposes. In some investigations. but they are rarer. if a file of pedophiles is available. then ID photos can be used to check whether people who work with children are in the file. These images may show the face of a suspect. Figure 5: facial recognition in a crowd. they cannot be used to successfully close investigations if the film only shows the top of the suspect’s head or if the images are blurred. Current automatic face extraction techniques work well with almost full frontal views of faces and when the quality of the video is good enough. facial recognition can also be used to interactively locate persons wanted by the police in video footage. looking for one person amongst a crowd of 100. then investigation by facial recognition may lead to success. If the recollections of the victim or the witnesses are precise enough to make a facial composite picture resembling the offender. the portraits must be extracted from the available evidence. The combined advances of video surveillance systems and facial recognition technology should enable more crimes to be solved using video data in the next few years. This application is subject to controversy. since it is often considered to infringe civil liberties. American and Australian customers have scored numerous hits with high-quality images. In any case. it is still possible to compare them with portraits of persons who are known to the police. Internet sites or copies of identity papers. By way of example. a witness’s camera. However. Even if the quality of the extracted portraits is highly variable. They may come from surveillance videos. hundreds of hours of video footage are analyzed and the “manual” search for excerpts in which faces are visible is a long and painstaking job. This would have a negative impact on the vigilance of control operators. it is not currently suited to cases in which a very small number of persons need to be identified in a crowd.1% false alarms. It is the reason why automatic assistance is necessary. Morpho’s experience in this field shows that these searches can already solve and correlate crimes. To begin with. Research is currently being made into the extraction of side and three-quarter views of faces. Another source of facial images is the facial composite picture.Criminal investigations and information Images are often made available for inquiries.000 passing people per day would operationally generate 100 false alarms per day. By way of example. In some cases. such as authentic or false ID documents or images posted on the Internet. 7 . Operational examples of the use of surveillance videos do exist. images of fraudulent use of ATMs or assaults close to an ATM can be used to solve crimes if the camera obtains well-lit. Our French. It is also interesting to note that certain criminal cases have been solved using low quality images. facial images. It can be used to search for precedents. Even if this technique were to reach the excellent accuracy level of 90% of persons actually found with just 0.

while fingerprints will always have criminal connotations and it is more difficult to acquire an image of an iris than of a face. As a consequence. Figure 7: biometric passport. the ICAO has recommended that the portrait should be the only compulsory biometric record. document holders are required to stand still in front of the camera. they must be cooperative and show their face. This check can be made by simply filming the holders when they present their documents. The operators who check passport control processes can also check any alerts received in response to these searches. With regard to travel documents. facial recognition can be used to check that they are indeed being used by their legitimate holders. Figure 6: facial recognition at airport. 8 . Control of identity documents Once the documents have been issued. but in the near future the check will be made as they pass through the checkpoint. In a non-criminal context.On the other hand. when travelers approach the border police for a travel document check using facial biometry. Today. Morpho developed a solution for this very purpose for the state of New South Wales in Australia. In this case. By way of example. but rarely include any other biometric data. it is quite easy to check the travel documents and make a comparison with the lists of wanted persons. it is quite normal to provide a photo. facial recognition applications can be used to guarantee that a single motorist cannot possess several driving licenses. interactive facial recognition is already possible in controlled passages. Morpho’s rapid and robust Face on the Fly* technology is capable of acquiring faces in three dimensions. By way of example. Civil applications. access controls and border controls Issuance of identity documents Facial recognition is particularly well suited to checks of the uniqueness of application for identity papers. By way of example. driving licenses include a photo of the holder. the SmartGates* for automatic passport checks deployed by Morpho have accelerated border formalities in Australian airports. without requiring the subject to stand still.

For example. The main advantage of portraits is that checks can still be made once the person has passed through the access barrier. Figure 8: facial recognition manages personal collections of photographs. since logins are possible if a photo of the user is shown instead of the user’s actual face. But thanks to portraits. But the system has come in for some criticism. if they already appear in older pictures in the collection. whose limits are still far from known. it is possible to check permanently that the people in protected zones are indeed entitled to be there. Products include iPhoto from Apple and Picasa from Google. then staff members could easily allow strangers to enter limited-access zones. which can be acquired without any special cooperation. it shows the full potential of facial recognition.This application is so easy to use that a broad range of usages is possible. Access control The purpose of access control is to check that anyone attempting to access a secure zone is entitled to do so. and future generations of biometric facial login systems will not be fooled by photos. They are very easy for the user if facial recognition is used. Facial recognition can therefore be used to extend access control by checking presence in particularly sensitive environments. 9 . Applications for the general public Access to computerized services Biometric facial logins are already possible on certain computers. Photo album management Facial recognition applications are now available to manage personal collections of photographs by showing the names of persons in photos. it would be easy for universities to check the identity of students when they arrive to take an exam (authentication checks). If the access control gates are unmanned. While this application may appear trivial. Recent algorithms are capable of detecting whether the face is indeed three-dimensional and mobile. Access controls are made in the same way as ID checks.

directions. etc.The Technology This chapter describes the different steps in the facial recognition process and describes the main technologies that are available today. Step 4: extraction of characteristics Most facial recognition algorithms use mathematical transformations in order to compare images. In this step. Step 5: representation as a template* and comparison A binary record. Step 2: face localization. The higher the score. is extracted from the transformed image. While some enhancements can be made automatically. These transformations can highlight the distinctive specific features of an image: frequencies. contours. By way of example. But it is much more difficult to extract a multitude of faces from a video taken outdoors. Figure 10: portrait comparison. In this step. the effects of ageing and expressions. inconsistent lighting can be corrected or unusable zones (masked by a veil. it is first necessary to find them in images that may contain all sorts of other information and adjust them to the same scale. the assistance of an operator may prove to be very useful when working on difficult images. Figure 9: Portrait acquisition. the higher the similarity with the image of the wanted face. the portrait can be re-acquired in order to obtain a better image that meets the criteria of the image assessment process. they need to be enhanced. because the precision of facial recognition hinges on the quality of the images acquired. The different steps of the automatic facial recognition process Step 1: image acquisition This step is decisive. scaling and alignment Before comparing faces. This step is quite simple when working on controlled portraits. automatic facial recognition systems may assess the quality of the acquired images. models can be applied to correct the orientation of the face. the effects of compression can be minimized. Transformed images can not usually be used by the operator’s naked eye. The comparator then compares this template with those of the images in the reference database and scores each image. or template. with the head positioned vertically. Step 3: enhancement of face images Once the faces have been found and calibrated. because each image only contains one face. In interactive acquisition. for example) can be detected and excluded. 10 .

learning algorithms. and machine Figure 11: bunch graph matching. i. This operation is called thresholding. It requires a similarity function that makes a clear distinction between ”hits” and false alarms. which imitate the analysis made by an operator. or 3D acquisitions. the lowest point of the chin or the details and color of the skin. The ideal case is when the operator only receives a candidate list when there’s a “hit” in the database. The operator can then make a decision and make changes to the system’s reference database* according to the requirements of the job in algorithms/. After detecting the landmarks of the face . Both categories of algorithms can be used for different types of face data: fixed images. when the quality of the query image and the images in the reference database are good. such as the corner of the eye. There are two prominent categories of algorithms when it comes to facial recognition: procedural algorithms. videos. For more details. which apply a mathematical logic in order to define and use the criteria that an operator may not be capable of interpreting.a process that may be manually assisted .e. thus preserving the operator’s resources and attention on relevant cases. visit http://www. They do this by using models designed to demonstrate how a face is distorted by its expression. The facial recognition algorithms This chapter contains an overview of the best known algorithms.face-rec. Procedural algorithms The main procedural algorithms use the visible facial landmarks. which is an excellent source of information. The most commonly used algorithms in this category are Elastic Bunch Graph Matching (EBGM) and the comparison of facial texture. the operator should only receive a small number of images that stand a high chance of matching the wanted person. These algorithms are used (at step 5 above) to convert images into templates in order to compare them. Facial texture analysis is used in particular to distinguish twins. and nothing otherwise. the middle of the upper lip.Step 6: thresholding and decision-making If the search is simple. 11 . orientation and lighting. age.the procedural algorithms attempt to measure the coherence between the parts of the two faces.

0486 0. Linear discriminant analysis (LDA). Examples include: • Vector projections.074 • Statistical methods.058 Figure 15: Sample neural network. These methods seek to measure the probability that a photo matches a statistical model of the face.0492 0. These networks use a set of cells that transform the information that they exchange with one another.0496 0.062 1. their respective probability and the probability of transition from one state to another. The network output is used to decide whether the face resembles the different faces in the reference database. Each face is represented in the comparator by its statistical model. 1.0488 0.0484 0.068 1. New facial images are represented as a sequence of successive Figure 13: Boundary detection with linear discriminant analysis.which is then projected in a smaller space. and leads to the training process. They resemble neurons.072 1.0494 = type A image = type B image Detection boundary I625 0. Y Y Independent component analysis (ICA). they must have close projected vectors. If the two original images match the same person. which separates different objects. It is the definition of the projection that is complex.g. There is a great number of training algorithms.the characteristics extracted in step 4 .070 1.060 1. Neural networks are defined by training. 12 . The most common methods are: Principal component analysis (PCA). including Kernels and SVMs* (Support Vector Machines). which extracts the most distinctive vectors from a space. Eigenfaces apply this principle. If they represent two different persons. Separation may be easier in higher dimensions Feature map separating hyperplane complex in low dimensions simple in higher dimensions Figure 14: Space transformation.0490 0. e. which keeps the axes as independent from one another as possible. synapses and nervous influx. DB I1 H1 25 x 25 Original image I3 O40 H50 I2 H2 O2 O1 X Figure 12: Axes created with principle component analysis. 0.Training algorithms Training methods rely on an abstract process in order to find independently an optimal organization based on examples. Non-linear methods. the characteristics extracted from the portrait are entered for input into the network.064 1.0482 1. The input used by these methods is a large vector . •N  eural networks.066 1. In the comparison phase.0498 0. then the projected vectors are more distant. by a number of states.

3D technology can be used. they generate probabilized detection and tracking hypotheses. which are consolidated in order to make a has demonstrated how the association of 2D data (texture) and 3D data (shape) improves precision compared with the use of only the texture or the shape. then its shape and the color or texture of every part of the face must be known. because they can be used to obtain more information about the face than a single view. The acquisition of three-dimensional images requires special sensors that are not yet widely available. Light sources The images used for facial recognition are usually taken in visible light. The actual comparison process uses a series of images of the same person that are sorted according to the quality of the views and the different positions that they represent.states. These sensors work by projecting structured light onto the 13 . At each step. This association significantly improves the robustness of the comparison of oriented faces. Special features of 3D Faces are naturally three-dimensional. Photos only contain the color of part of the face. It is called Face on the Fly* technology. The European 3D Face project (http://www. such as terahertz*. But infrared images can also be used and research is currently looking into other data types. information. but they can also be used to locate faces in images (step 2) and to extract visible facial landmarks. It can be viewed from one video frame* to the next. A range of tracking techniques has therefore been developed. Using all the information of the face can only serve to improve the precision of facial recognition. Using these different views rather than a single view – even if its quality is superior – will always improve the precision of the search. These multiple views are useful for facial recognition purposes. Figure 16: examples of pose angles. For each face in the reference database. Training algorithms are obviously used in step 5 (representation as templates and comparison). Special features of video processing A person’s face appears several times in a video. If a face is to be completely represented. Morpho has developed an innovative 3D acquisition concept based on stereoscopy that is capable of acquiring a face on the fly when a person passes through a control gate without stopping. the level of probability of the sequence can be determined in order to decide whether the resemblance between the two images is high or not. or 2D. The most robust techniques use movement statistics models. even without any 3D sensors. Therefore they only contain partial. 3D morphable models can be used to take one or more images of the same face and associate a 3D shape that matches the face as closely as possible. face or by stereoscopy*. Traditional face comparison techniques rely on these incomplete data.

which measures the percentage of fraudsters who are mistakenly accepted. while out of 1. the probability that the system will recognize him drops a little. If the same person allows his beard to grow and pulls a face. the more secure is the application. remains a long way away from the camera and conceals a large part of one of the sides of his face with his hand). The tests in NIST’s FRVT 2006 demonstrated that. weak controls Very low false acceptance: security FALSE REJECTIONS Figure 17: False acceptance – false rejection graph. with frontal portraits taken in a strictly controlled environment. changes of hair. with high resolution and only slight differences in age. resolution and lighting. Figure 18: Variations with age and orientation. but still remains high. for example. The lower the rate of false acceptances. then the Very low false rejections: comfort for users. spectacles. But we all know that the ideal conditions implemented by the NIST for FRVT 2006 are difficult to achieve operationally. But in reality. does not face the camera. For example. changes in weight. 999 would be detected by the automatic control system. in the ideal passport control application. Ideally. pulls a face. ageing.000 authentic passport holders. facial recognition is also confronted with problems due to physical changes and changes in appearance: expression. injuries. 14 .000 fraudsters carrying passports that do not belong to them. only 10 would not get through the automatic check and would have to call on an operator. if the person allows his beard to grow. then the automatic portrait recognition system will recognize him with almost the same reliability. FALSE ACCEPTANCES with 0. out of 1. This means that. biometric systems can be characterized by the graphs on the model below. the greater the comfort for users and the more limited the work done by operators. if a person allows his beard to grow. illness. But if a number of small changes are accumulated (by way of example. as if his beard had not changed. the false acceptance and rejection rates should be zero. whereas they are not fraudsters. Figure 19: Variations with accessories. False rejections totaled just 1%. The current facial recognition algorithms can only tolerate limited variations in the portrait. • the false rejection rate (FRR*).1% of false acceptances. In addition to orientation. The lower the rate of false rejections. etc. hats. facial recognition is a very precise biometric technique.Performance… and comparison with other forms of biometrics Two figures are required to measure the performance of a biometric technique: • the false acceptance rate (FAR*). which measures the percentage of persons who are not accepted.

and since these criteria are not independent. Since there are no test samples that can be used to measure the impact. with the appropriate statistical validity. Nevertheless. the drop in the precision of facial recognition is not predictable in absolute terms.probability that he will be recognized is much lower. and will be more difficult to adapt to the working procedures of the operators. depending on their own operational needs. Easy to represent in 2D Intrinsically 3D Distortions Pupil dilation Resolution Requires high resolution Distortion limited by the elasticity of the skin Usually standard 500 dpi Strong command of the identification process by generations of fingerprinting experts Highly variable. Comparison of biometric methods Iris Fingerprints Face Two persons may resemble one another very closely One face per person Uniqueness Every iris is unique Every fingerprint is unique Number of images Two irises per person to be acquired 10 fingerprints per person. according to expressions Acquired at any scale Underdeveloped expertise Maturity of the processes Underdeveloped expertise associated with the technology 15 . A very slow system will necessarily find fewer hits than a fast system. Morpho advises every potential user to proceed with tests and measurements of their own data. But these tests must not simply take account of the purely algorithmic precision. If one compares facial recognition with fingerprinting and iris recognition. Consequently. state of health. Facial recognition already works well and still has plenty of potential for improvement. plus other parts of the body where friction ridges are located Stability over time Invariable from birth Invariable from childhood Representation in the dimensions Easy to represent in 2D and few problems due to orientation Changes with age. it is quite improbable that it will achieve the same levels of precision as iris or fingerprint recognition in the short term. etc. of the different criteria that minimize the precision of facial recognition. then it becomes clear that facial recognition is intrinsically “trickier” than other biometric techniques. They must also take the speed of execution of the algorithms into consideration. with the same power and precision.

which are specific to facial recognition. But the operators only received false alerts. For example. The unsuccessful experiment in Tampa. it is advisable that every potential user of facial recognition conducts tests in order to assess the suitability of the technology to their operational Integration with databases allows for a simplified recovery of existing data and allows legacy data acquisition processes to be used. without impacting the potentially complex processes that are used with the civil status register. A facial recognition system was deployed in order to recognize wanted persons in a crowd. processes and the benefits of using a number of methods that increase the performance of searches and require operators to verify only the most difficult cases. Most potential users of facial recognition therefore already possess a number of information systems that can be interfaced with a facial recognition system: • databases of individuals (with portraits) or criminal cases • other systems designed to compare biometric data • biometric data acquisition systems There is almost always a business interest to be gained by integrating a facial recognition system with the existing environment. Morpho has developed generic interfaces that meet the ANSI/NISTITL 1-2007 standard for the data format for fingerprints. then it is highly likely that the shoplifting offense will also be solved. in order to add a facial recognition function to a civil status register in order to detect identity fraud. Integration with other biometric systems provides for both optimized.Implementing a facial recognition system The deployment of a facial recognition system must take both technical and human factors into consideration. faces and other biometric data. • it can make more extensive. multi-biometrics does increase performance: • it allows cases to be processed. if it is known that a first offense of bank card fraud and a second offense of shoplifting were committed by the same person. and consequently more efficient links. By way of example. Therefore. in a criminal police system. when one modality is absent or is of poor quality. Florida. because the images show the same face. USA in 2003 is one notable example. Integrating facial recognition in the existing technical environment Photos are a very common form of data. Even if biometric searches are not natively consolidated. redundancy-free data acquisition 16 . it is always preferable to keep the same civil status register and operate the facial recognition system in back office mode. Figure 20: identification verification screen within the verification application. Two isolated systems would not come to the same conclusion. Preliminary tests in a controlled environment and an elementary probability calculation could have concluded that the technology of the time was not suited to facial recognition in crowds. are quite important. In an effort to facilitate the integration of facial recognition with other information systems. Managing the expectations of customers and operators Facial recognition seems to be so simple and intuitive that the expectations relative to this technique may be out of all proportion. Some of these factors. and if the bank card fraud is solved using fingerprints.

even little children are capable of recognizing a face. Research reports explain perfectly the state of the art.Figure 21: on-the-fly portrait acquisition for traveller screening in airports. Expertise in checking searches Visually recognizing people may appear to be simple. Reactions range from total hostility (against a society that some people feel is obsessed with security) to a certain amusement at being a pioneer in the use of new technology. When acquiring an image in the field for an identity check. It is also important to cooperate with academic researchers and facial recognition technology vendors in order to make progress in terms of both the practices and the tools in this field. When working on subjects that one knows well. Morpho can provide support to its potential customers in this assessment process. the conclusions and the outlook. Morpho can provide training in its products. filming people behaving violently in a demonstration or choosing which images from many images of the same face should be used for search purposes. (1) BKA. Morpho has developed a very simple pilot system containing the most recent advances in algorithms that can be installed and programmed in less than one day. visual recognition becomes a tricky task. the tests conducted. In order to conduct these tests. it is impossible to support the users one by one. The criteria that measure this quality are defined by the ISO* standard. And when the person is unknown. Developing the expertise of operators Acquiring facial images and recognizing people on photos is not easy and requires the development of specific expertise. The relevance of recognition criteria (e. Bundeskriminalamt. In applications for the general public. but they are not as robust as the methods used to check fingerprints. or if they have an unusual expression.g. The independent tests conducted in 2006 by the BKA(1) and also the NPIA(2) serve as an example. Similarly. This is the reason why close attention must be paid to the ease of use of biometric tools for end users. they are all novices. the operator is the person who is best placed to judge the suitability of an image for facial recognition searches. and it is essential to give them clear and concise instructions on the behavior to be adopted. the stability of wrinkles) remains to be scientifically established. check the results that may be obtained for a given target application and measure the workload required to obtain these results. and that are seen face to face rather than on a photo. Whether they be cooperative or hostile. But the judgment of an operator remains the best guarantee of good image acquisition. resulting in many errors detected by academic studies. applications before proceeding with deployment. Therefore. the operator’s expertise is decisive in making optimal use of facial recognition. But even known persons can be misidentified if they appear on blurred or old photos. An organization tasked with making technological recommendations to the British police force in order to improve efficiency. Helping the end users End users react very differently to biometric systems. such as passport controls. The SmartGate* passport control system deployed in Australia meets this need. The end users are happy with the system and prefer automatic passport controls using facial recognition to conventional control gates. the quality of the photos is variable and the angles and lighting differ. In this way. it is possible to test facial recognition usage scenarios. (2)  17 . A number of methods have been developed to improve the visual recognition of persons. it is necessary to develop methods and training for operators in the use of the facial recognition system. NPIA: National Police Improvement Agency. in order to avoid making mistakes and wasting time. the German police authorities. they must be as enjoyable to operate as possible. and the automatic acquisition systems are capable of verifying most of these criteria. Expertise in data acquisition If facial recognition is to be efficient. then the quality of the images in the reference database must be satisfactory. Automatic checks cannot be 100% reliable and the subject that is filmed may have particular physical properties that prevent certain criteria from being reached.

2007. ranging from technology to politics. for static images and video Development and integration of FR utilization in vertical markets Deployement or upgrade of integrated FR oriented solutions in public security market Figure 23: Facial recognition: value chain. 2008). etc. applications and even standards. the facial recognition market should exceed € 622 million by 2011 and €1 billion in 2013. It has resulted in the creation of a reference database of high quality portraits. Development SW Development System Integration 1. facial recognition accounts for 16% of the market (source: Frost & Sullivan. in 2D and 3D.Photo-video Development of high accuracy face recognition algorithms. The driving forces 16% 84% • The creation of an international standard (ICAO) for travel documents. Figure 22: Face recognition share in the biometric market. 18 . Totaling €72. iris recognition. the value of this market is growing constantly. Also according to Frost & Sullivan. with an average annual growth rate of 54%.7 million in 2006 and about €250 million in 2009.). Facial Recognition: value chain Components / Imaging sensors Face Tech. This standard specifies three possible forms of biometrics: iris recognition.The market associated with automatic facial recognition The market for facial recognition Compared with other biometric techniques (fingerprinting. Driving forces and obstacles A number of different factors drive or hinder the development of facial recognition. These estimates correspond to all the links in the value chain of developments in the field of facial recognition.2D/3D cameras 2. fingerprinting and facial recognition.

governmental or inter-governmental identity programs (electronic passports. driving licenses. Notably. Facial recognition of good quality images is a mature biometric technique and the algorithms continue to progress. These implementations have been a real success and the expectations of potential users are becoming more precise and realistic. can solve difficult cases in which the data from one of the biometric techniques is of poor quality and can correlate the connections made by different biometric techniques. which allow for more precise facial recognition. etc. • Facial recognition is often quoted as an infringement of civil liberties. 19 . • Existing implementations. The obstacles • Facial recognition is intrinsically more difficult than other major biometric techniques. • The proven benefits of multi-biometrics. Multibiometrics cut staffing costs in biometric research. opening the way for the combination of 2D and 3D facial recognition techniques.• The highly favorable conclusions of FRVT 2006. • A pplications are broader than for other biometric techniques and the potential new users must be introduced to the field of biometrics and understand its benefits. ID cards.) and automatic border passport control programs. • T he availability of video cameras with improved resolution. These cameras take better pictures.

The University of Lausanne.The experience of Morpho in the field of facial recognition Morpho has been active in the field of facial recognition since 2002. successfully developed a strategy for the use of Morpho’s facial recognition for demonstrations. In 2008. It also surmised that the combined use of Morpho’s facial recognition and fingerprinting technology could allow subjects to be identified without calling on fingerprinting experts.000 people had passed through a SmartGate. 20 . 150. Texas in May 2008. Morpho has significantly improved the technology. The Australian police force concluded that facial recognition offered decisive benefits for the correlation of criminal cases and now uses Morpho technology in operational applications. By the end of 2008. the Mexican police confirmed that Morpho’s technology was the fastest and the most accurate on the market in operational tests. • Morpho developed the SmartGate border control system that is now used in all international airports in Australia and will soon be deployed in New Zealand. This pilot allowed a number of conclusions to be drawn: The Morpho subsidiary MorphoTrak received the “Best Biometric Identification Technology” award for its facial recognition pilot system when it was exhibited at the Global Border Security Conference and Expo in Austin. even with significant differences in age. demonstrated how Morpho’s facial recognition reached a level of precision higher than 94% when identifying individuals in its collections. in cooperation with the Romande regional police force in Switzerland. Some of the most significant milestones in our activity include: • 2002 assessment of the algorithms on the market: we opted to start with technology from Cognitec. This project uses the photos on biometric passports to speed up and facilitate passport controls.000 faces. Since then. The tool allowed numerous police agencies to test the possible applications of facial recognition. the Paris police authorities helped investigators by piloting a Morpho facial recognition system on a reference database of 470. The Pierce County Sheriff’s Office in Washington. which was found to be the best on the market at the time by the NIST’s FRVT 2002. USA. • Morpho deployed a pilot tool designed to assess the technology with simple and efficient user interfaces. In 2009.

Major advances are continuously being made in terms of both the quality and diversity of the algorithms and the development of dedicated products adapted to the needs of the market. Morpho finished first in the NIST’s portal challenge in March 2009.3dface. but also the optimization of portrait acquisition and the extraction of portraits from surveillance video footage. to help the police to solve criminal cases or simply for identification purposes in civil applications. • Morpho’s research laboratories continuously develop and improve the basic technology.php http://www. including the 3D Face project. Websites: http://face.hcfdc. we are listed or quoted under the name of Sagem Sécurité. Morpho is making significant investments in innovation and the development of its facial recognition technology in order to consolidate its position as leader in biometrics. This product is designed to perform identification tasks. which has concluded that the use of both 2D and 3D facial data improves the precision of facial recognition for travel document applications (http://www. benchmarks having taken place before this date. Morpho takes part in cooperative research projects in the United States and NB : On May 27 2010. Sagem Sécurité changed its name to Morpho. For• Morpho also developed MorphoFace™ Investigate* on the basis of recommendations made by our users working in dedicated Focus Groups. org/). 21 . Morpho received the HCFDC (French High Committee for Civil Defense) innovation trophy for its Face on the fly face acquisition technology. Research focuses on facial recognition http://www. As a result of this research work. In May 2009.

Face On The Fly is an innovative technology developed by Morpho.An image extracted from a video recording or stream. Its mission is to contribute to the development of standards used to standardize international air transport ISO . SVM .Glossary and acronyms A Alert .A database of persons of interest to be identified in the images and videos to be processed. An international organization that is part of the United Nations.Face Recognition Vendor Test. The purpose of this technology is to acquire facial images when a person passes through a control gate. The suspect is presented to the witness amongst a group of physically similar people. Witnesses must then decide whether they recognize one of the members of the group. L Line-up .MorphoFace™ Investigate.International Organization for Standardization. without stopping and without having to look at a particular camera. False acceptances and false rejections are used to measure the performance of a biometric system. T Template . N NIST . FR . S SmartGate . Frame .Multi Biometric Grand Challenge (http://face.The code extracted from an image of a face by image processing.International Civil Aviation Organization.A type of biometrics. A test organized by the NIST to make advances in research into the recognition of persons from a distance using facial recognition and iris recognition.False Acceptance. Face on the Fly . Alerts are generated when the recognition algorithms conclude that a person seen in the video is very probably contained in the reference database. A prominent technique used to solve classification problems. This system is essentially designed for use in police investigations. Facial comparison is carried out via extracted templates. 22 . I ICAO . Several images are acquired by a series of cameras to create a three dimensional view of the face.Stereoscopy refers to all the techniques used to reproduce a perception of a contour from several flat images. R Reference database . A facial recognition system developed by Morpho.An electromagnetic wave in the electromagnetic spectrum between infrared (the optical domain) and microwaves (the electronic domain). Stereoscopy . False acceptances correspond to fraudsters that the biometric control fails to detect.False Vector Machine. The purpose of the project is to speed up customs clearance in Australia’s international airports using the portraits on electronic passports and facial recognition.nist. Face On The Fly technology won the French High Committee for Civil Defense’s award for technological innovation in May 2009. False rejections correspond to authorized users that the biometric control fails to recognize.An alert is an automatic and interactive event that is generated when analyzing video streams.National Institute of Standards a n d Te c h n o l o g y. Terahertz .Line-up techniques are used by the police to determine whether a witness has spotted a suspect. Alerts appear in the user interface by displaying the image of the person spotted in the video next to the image in the reference database. A frontal projection of this image can then be compared and used for authentication or identification purposes. F FA . Morpho deploys a system in Australian airports as part of this project. An international organization made up of the national standardization institutes from more than 100 countries. Modality .SmartGate is a project run by Australian customs. M MBGC . It can be used to solve cases from portrait traces left on the scene of the crime. T h e A m e r i c a n standardization organization and a member of ISO. MFI . FRVT .

Advances in Biometrics. Forensic Art and Illustration Karen T. ISO/IEC 19794-5. Federal Bureau of Investigation Forensic Audio. Kleinber. “World Biometrics Market”. 2005. Arnout C. L Sirovich and M. Frost & Sullivan (2007). Turk and A. ISBN 0849381185. Vanezis. BKA. Germany. “Quality Standards for Forensic Opinions on the Identity of Living Offenders in Pictures”. Ph. 2007. Institut fur Humangenetik und Anthropologie Universitatsklinikum.html Automatic Face Recognition. “Portrait parlé”. Kirby. Katherine Jo Strandburg. www. 1987. “Morphological Classification of Facial Features in Adult Caucasian Males” Vanezis. Biometric Data Interchange Format . ANSI / NIST: Data Format for the Interchange of Fingerprint. Ulm. 9th Biennial Scientific Meeting of the International Association for Craniofacial Identification. October 2006. and Burton . Jenkins.police. . “Face recognition using eigenfaces”. 2007. 1991. Goldsmiths. Justice Of The Peace Volume 172. CRC Press. Becker. page 146. Foto-Fahndung. Video and Image Analysis Unit. “Photo identification: facial metrical and morphological features in South African males”. London. Gesichtserkennung als Fahndungshilfsmittel. Journal of Cognitive Neuroscience. www. A. 1996. and international inventory”. M. 3(1). Laan van Ypenburg 6.bka. “Facial comparison ‘area of expertise’: formulation of the issues. Steyn. UK. Spaun. applications within law enforcement. UK. Privacy and technologies of identity: a cross-disciplinary conversation. Market and technology review.Forensic Science International. Daniela Stan Raicu. June 2009. 9780387260501. Abschlussbericht. July 2007. M. 2008. Etude du signalement descriptif. Journal of the Optical Society of America A 4: 519–524. Pentland. University of London. et al. Proc.D. Burton. Pharm. Facial. 1991. Identité judiciaire. “Facial Comparisons by Subject Matter Experts: Their Role in “Failure of Anthropometry as a Facial Identification Technique Using High-Quality Photographs”.Part 5: Face Image Data. Roelofse.See also: http://www. faces”. Turk.bka.pdf “CCTV on trials”. Taylor. Ambika Suman. Rosing. Report.npia. & Other Biometric Information. M. Josh P Davis and Tim Valentine.pdf . “Eigenfaces for recognition”. de/kriminalwissenschaften/fotofahndung/faq.Bibliography ForshungsProjekt.Journal of Forensic Science. Department of Psychology. Ruifrok. Préfecture de Police de Paris. R. . .de/kriminalwissenschaften/fotofahndung/pdf/ fotofahndung_abschlussbericht. University of Glasgow. “Low-dimensional procedure for the characterization of human “Limitations in Facial Identification: The Evidence” . 2000. F. Biometrics and Their Training”. Glasgow. 9780849381188. Nicole A. 2497 GB Den Haag. ISBN 0387260501. IEEE Conference on Computer Vision and Pattern Recognition: 586–591.Journal of Forensic Sciences. January 2008. Springer.W.

rue Leblanc .Polyvalence.F-75512 PARIS CEDEX 15 .075 e 440 305 282 RCS PARIS .Photos : © Getty Images / © Fotolia / © PhotoAlto.morpho.D-10156 Phone: +33 (0) 58 11 88 76 Registered Office: Le Ponant de Paris .com .Fax: +33 (0) 58 11 87 81 www. Printed on paper made from responsibly managed forests .France Société anonyme au capital de 159.

Sign up to vote on this title
UsefulNot useful