基于关键帧信息融合的视频人脸识别方法研究-电子与通信工程专业论文.docxVIP

ABSTRACT In the procedure of video-based face perception technology, the problem of how to steadily and exactly recognize faces which have substantial changes in posture and scale in video is inevitable. In such situation, the traditional face recognition algorithm only recognize faces with a single static video frame, depending on the frame we have chosen. Obviously, this way is neither reliable nor accurate. Even though for the sake of improving system stability, adding up several video frame sequences to identify results using mathematical methods and finally work out a conclusion, there is often more harm than good because of the increasing computation complexity. As a newly developing topic, Compressed Sensing (CS as an abbreviation) provides solution seeking mindset in numerous fields. Although face recognition algorithms based on CS always aim at static image at present, considering the direct information sampling characteristic of CS and the frame-to-frame correlation feature of video, we take both into account. In detail, randomizing the video frames’ information in advance, then use it in face recognition based on CS in order not to exceed system computational burden too much, but get a better system performance. By studying the face perception technology, especially the video face recognition currently, this paper puts forward a video-based face recognition system using keyframes information integration in the light upon understanding and comparing several face recognition arithmetic. Firstly, extracting keyframes from video sequences which are suitable for face identification in accordance to the similarities of color histogram. Then Adaboost+Haar algorithm is used to detect the face in keyframes, followed by denoising and normalizing the detected face region images. Finally, with the knowledge of Compressed Sensing theory, proposing a CS method, which is appropriate for video-based face recognition on the basis of keyframe information integration that r