Challenges
Generic people identification systems are able to establish corresponding between images and make decisions on person's identity using exact frontal images of a person. But in many cases images from surveillance cameras, used for identification are not frontal, as they are made in real-life conditions and capture a person in movement. In a word it is next to impossible to get a frontal image to be used for identification. In these cases generic identity verification systems have rather a high false decision ratio or simply fail to take a decision.

Solution
Trying to increase the efficiency of identification technologies IDOT has developed a cutting-edge face reconstruction algorithm. Using several non-frontal images the solution creates a 3D face model and generates an exact frontal face image. This image is then used to establish the correspondence between an image in database and the image captured by surveillance camera to determine person's identity. This technology significantly improves the reliability of existing 2D identification systems and increases security level, eliminating false alarms and extensive human supervision.

Challenges
The international ISO/IEC 19794-5 standard defines conditions for taking photographs for face image data. All pictures for modern biometrical passports must be compliant with this standard. However working with huge digital imagery data in photo studios usually requires lots of attention and consumes much time, when you have to pick up an image, fully corresponding with the standard requirements.

Solution
IDOT's solution for quality control of human face digital imagery solves this problem easily. It analyzes the digital image from a camera and detects whether the image meets the ISO/IEC19794-5 standard requirements, for example whether the image is exactly frontal, the eyes are open and there is no hair across the face. Thus even a not qualified operator can quickly pick up appropriate images.

Challenges
The need in exact person's identification is enormous:

• At entry points, for example at passport control counters in airports or on railway stations, by access to a security zone, etc. ;
• In enterprise control applications for working time monitoring;
• In financial security applications: bank employees need to be sure that the customer applying for a huge loan is not included in a criminals' or a black list database.
• By identifying wanted persons based on the video data from surveillance cameras in public areas.

The decision about person's identity must be fast and exact. Existing face detection and analysis systems haven't found vast acceptance, as they require usually much investment in equipment and extensive training for security employees. So until now almost no automation technologies are used.

Solution
The solution by IDOT is a precise answer to security and efficiency challenges. The deployment of automation solution in many cases does not require installation of new expensive equipment. Our software solution, a frame grabber and CCD camera turn a generic PC into a high quality identity verification system. The automatically captured face image is compared with the images in a database (in access management or security systems) or with the passport image (at passport control counters) and the decision about correspondence is met. With this solution security employees receive a powerful automation tool, which gives them more confidence and prevents inattention-mistakes, possible during long shifts and overtime work.