Face Of The Video Image Compression Encoding

The compression and coding of facial image and video are one of the most important topics in the field of digital image processing techniques. It is widely used in digital communications and web services. The problem is subdivided into three parts: accurate face detection algorithm, real-time face tracking algorithm and efficient face coding algorithm.The main contribution of this paper lies in four aspects as below:1) To detect faces in the video or image sequences, an algorithm based on temporal difference and background subtraction is represented. According to the elliptic shape feature, the location of face can be calculated by least square approximation.2) Kalman filter is introduced to predict the location of faces in the next frame of images. Then the region of interest(ROI) is reduced to a much smaller area. Apply the algorithm in 1), we can get a real-time face tracking algorithm.3) Partition Principal Component Analysis is proposed in the compression of face images. Aiming at the reduction of time complexity, PCA is first converted to Singular Value Decomposition problem. Then an Eigenspace Update Algorithm is designed to adapt to real-time requirement.4) Using Microsoft VFW library, a face video coding & decoding system is developed. The system incorporates many functions in a unit, including image capture, image display, coding and decoding, etc.