Algorithms Research And Hardware Implementation Of Face Detection And Recognition

With the ever-accelerating process of the rapid development of computer networktechnology and information technology, information security and public safety has been moreand more concerned. Accurate recognition and identification is an important prerequisite toensure system security, therefore face recognition technology is welcomed by more and morepeople. It is an identification technology based on biometrics, is a multi-disciplinary Edgeapplication technology which crosses the image processing, pattern recognition, computervision, biology and other disciplines.This article carries out research and analysis of the commonly used algorithm of facedetection technology at home and abroad. Based on the comprehensive analysis andcomparison of these algorithms: At the aspect of face detection, do camera collected facedetection using improved Adaboost algorithm and pre-treat the face images by the ways, suchas high light of face images, gray-scale transformation, sharpening, median filtering, contrastenhancement, Image binarization pretreatment in order to establish the face database; in facerecognition aspect, analysis and test the ORL face database and self-built face database usingprincipal component analysis method (PCA), line analysis method(LDA) and theimprovement and integration algorithm method, The result of simulation experiments showsthat the improved algorithm not only has a good detection and recognition performance, andthe algorithm has low complexity and advantages of ease hardware implementation.In hardware implementation aspect, the algorithm code is ported to the TMS-DM6446devolopment board, and optimized.The experimental results show that: The system has abasic face image acquisition, preprocessing, detection and identification capabilities.