Improvement and Application of BMA in Facial Feature Extraction Based on MPEG-4

The block match algorithm (BMA) can efficiently extract facial definition parameter (FDP) and facial animation parameter (FAP) in human face video which is based on MPEG-4.Firstly,video frequency is analyzed and video illumination is smoothed so as to reduce luminance mutation.Then the block match algorithm is used to do the match of picture element.For the purpose of decreasing computation,the limitation for the position and gradient relationship of those feature points is appended.Furthermore,in order to get an exact match result,the correlation coefficient is improved.The effect of improved algorithm for experimental results is prominent