Specific Person Detection And Dased On Elastic Graph Matching Metchod Of Software And Hardware Implementation

With the rapid development of biometric recognition technology,face recognition teclmology is more and more used in human life,such as security system,human search system,access control system,etc.How to detect specific people accurately an d quickly from the image through the face has important practical application value.Firstly,the current situation of face detection and recognition is analyzed.On this basis,a method of face recognition based on Gabor elastogram matching method is given,and the software program and hardware structure of MATLAB for specific person detection are designed.Th e research results of this paper can be divided into two parts:(1)Firstly,this paper studies the algorithm of specific person detection and its realization in MATLAB.In order to remove the interference of background color similar to skin color,a rough location method of face detection based on YCgCr and YCgCb color space is presented.Secondly,the face recognition algorithm based on Gabor feature is improved,and the optimal weight parameters of feature points are determined through experiments.The image database of specific person detection is constructed,and the Mtalab program is written to realize the detection of specific person.The test results show that the recognition rate of specific person can reach 96%.(2)Based on the realization of the algorithm software,the hardware structure of the specific person detection Is designed.The whole hardware system includes skin color coarse location modale,face detection and recognition module,storage module and so on.Verilog HDL code of each module is written and simulated with Modelsim.Finally,the whole system is integrated into ALTERA STRATIX V device and verified on the FPGrA board.The results show that the hardware structure of the paper is correct.