面向视频的人脸识别算法研究与系统设计-软件工程专业论文.docxVIP

优秀毕业论文 精品参考文献资料 ABSTR^CTAB ABSTR^CT AB STRACT With the rapid development of artificial intelligence and computer vision，because of stability and difference advantage，biometric(fingerprint，face，iris，signature) recognition be widely used in automatic identity recognition，and achieved good effect of certification．Among them，face recognition has the characteristics of nature， convenience and non contact，SO it has become a hot spot in biometric identification． More and more domestic Internet，IT companies aimed at the face recognition of the internet finance business opportunities，launched its own face recognition products． Users who using these products，can be very convenient tO carry out the registration， account，payment and other operations．However,when the face recognition is used in video surveillance，access control systems and other practical applications，due to the influence of the external factors such as posture and illumination，occlusion，can not meet the requirements of people．Therefore，in face recognition，the study of that how to reduce the impact of external conditions，how to effectively represent the face and other aspects has important significance． Face preprocessing，face recognition and face recognition system are mainly studied in this paper．In order to get a better face image in complex conditions，this paper combines 4 methods to achieve face preprocessing．At the same time，this paper studies and analyzes the LBP and LTP operators，and puts forward LMTP to supplement the description of texture features．Finally,the face detection，preprocessing and recognition algorithms are applied to the video face recognition system． The main work of the thesis iS as follows： 1．Research on face image preprocessing algorithm，analysis the advantages and disadvantages of different algorithms，and puts forward the preprocessing algorithm which is based the four steps of gray stretch，Gamma correction，Gauss of difference and contrast equalization，this preprocessing algorithm