Overseer FaceCode

Face recognition system Overseer Face Code is a software complex designed for automatic stream face recognition and saving received information to the database. Intellectual system kernel is based on the latest neural network algorithms that guarantee the highest recognition level.

Face recognition system Overseer Face Code provides automatic image processing, search, capture, and comparison with the database for its identification. Overseer Face Code implements two functions: face detection function that helps to create a chronological archive of the faces detailed from the general background, and recognition function itself.

Overseer Face Code is able to work with video data stream from the camera (e.g. video capture boards, logger) as well search among the existing video files (e.g. security system archives).

Application areas:

• person control services (passport, frontier);
  
• identification control services (document, database, archive, internet access);
  
• working time control systems (checkpoints);
  
• physical access control systems (offices, banks, enterprises, guarded territories);
  
• person investigation in the public places (streets, squares, railway stations, airports, stadiums, shops, casinos);
  
• archive photos person identification systems;
  
• safes, bank deposit cells;
  
• ATMs, POS-terminals;
  
• security and alarm systems.
  

System service functions:

• parametric face capture and identification in the frame range, with different priority levels;
  
• captured faces visualization;
  
• dynamic portrait gallery display;
  
• keeping the recognized faces database;
  
• "unfair list" maintenance, alarm on the "suspicious" person being in the frame;
  
• portraits search and display by different criteria (date, time, number of people, camera number, person name);
  
• picture export to the graphic file of exchange buffer.
  

Technical features:

(Setting the system technical features is based on the results of testing the recognition algorithm using BioID (Switzerland) and FERET (USA) face databases)

• time of face searching in the frame: < 40 ms;
  
• correct face capture coefficient - over 95% (BioID);
  
• time of the face identification by comparison with 1000 records database - under 50 ms (FERET);
  
• correct identification coefficient - over 95% (FERET);
  
• probability of the faulty access granting (FAR) - 1%;
  
• probability of the faulty access denial (FRR) - 4.0 %.
  

Picture quality requirements:

• Position: Face position must be as much frontal as possible. ±150 head slope is acceptable, as well as head turn in the ranges of -100..+100 from initial frontal position.
  
• Illumination: It is recommended to set illumination so that the face is evenly illuminated, without any glares and flashes.
  
• Distance between the face and the camera must be sufficient for the horizontal face dimension to make from 20 to 95 percents of the frame width.
  
• Minimal face size must be not less than 72х72 pixels with the resolution over 300 dpi.
  

Camera requirements:

• Analogue, resolution - over 540 TVL, having rear illumination.
  

System requirements:

• Operation system: Windows 2000/XP;
  
• Processor Pentium IV 1,5GHz of higher;
  
• RAM over 512 МB.
  

To keep an archive, Overseer Face Code can use different databases as the system has a separate reprogrammable unit to solve these problems. In case if unknown databases are used, a technical description has to be provided by the developer.

Unit structure allows to deliver Overseer Face Code as a ready solution as well as a separate unit. Due to an interface for interaction with different software products Overseer Auto Code is easy to integrate into the ready video surveillance systems.