Overview:

IIT Face Detection and Tracing Software Development Kit (IIT FDT SDK) - software solution providing a set of functions for Face Detection, Face Position Determination and real-time Face Tracing. Face Detection procedure extracts human faces on color or grayscale close-to-frontal images and returns bounding rectangles for all faces detected. Face Tracing procedure allows accumulating and storing the information about the human face (different views and mimics of the same face) until this face will be turned away or quit the controlled zone.

This solution is created to be a compatible and flexible part of different client-oriented biometric systems.

Data:

• Video input:
  • Digital images or video flow with resolution not less than 320х240 pixels
• Detection output:
  • Number and position (list of bounding rectangles) for all detected faces (for a single frame)
  • Number and position (list of bounding rectangles) for all traced faces (for a sequence of frames)

Features:

• Image format: Minimal face size - 60x80 pixels.
• Operation modes:
  • Face Detection;
  • Face Tracing.
• Face Detection rate: up to 0.95
• Probability of face loosing during the tracing: down to 0.05
• Performance: up to 100 fps for 320x240 image for Intel Pentium IV 2,4GHz

Advantages:

• Non-critical to video hardware. Low cost USB camera is compatible for successful operation
• Video sequence registration and storage for each face. More flexible forming of "secure events" by "entering" and "quieting" of human face in controlled zone
• Convenient API. Easy incorporation of face detection and tracing component into the user's applications