Research Of Face Detection And Recognition Based On Adavanced Adaboost And LDP Algorithm

As an emerging biological characteristics identification technologies, the facerecognition technology refers to pattern recognition, image processing and computer visionand other related areas. It has important academic research value and broad applicationprospects. This paper mainly focuses on the face detection and identification based onimproved Adaboost and LDP algorithm. The main research achievements include:(i)A face detection method based on skin segmentation and Adaboost algorithm.Using the YCbCr color space to carry out the segmentation operation of human skin, filtersout the candidate face regions which correspond to the distribution of human face forpositioning face areas initially. Then locates the face regions further using the improvedAdaboost algorithm. The experimental results show that the proposed method not onlyaccelerates the speed of face detection, but also improves the detection rate.(ii)A face recognition method based on improved LDP features. The methodnormalizes the main direction of the LDP encoding to deal with the rotation of image andreduces the dimension of facial feature vector. At the same time, the method extracts thestructure contrast information to weight the face LDP blocks to enhance the robustness ofthe facial feature. The experimental results show that the improved algorithm can not onlyimprove the recognition rate, but also accelerate the speed of identification.(iii)A face recognition system based on video stream. By integrating the facedetection and recognition algorithms proposed in the previous chapters，this paper designsand realizes a face recognition system based on video stream, which lays the foundationfor further research.