Research And Implementation Of Video Summarization And Image Detection Technology On Mobile Platform

Nowadays computer technology and mobile Internet develop dramatically, and there are more and more multimedia applications on mobile platform emerging. Since mobile device has limited bandwidth and screen size, it requires more humanized and intelligent mobile terminal for image and video processing. Therefore, video summarization technology emerges. This thesis focuses on research of a video summarization system on mobile terminal which takes keyframe as the main summarization form and has a good user experience. Contents of this thesis are shown as followed.Firstly, the thesis introduces related technology background, including a brief introduction of video summarization and the current status of research. It introduces the two basic forms of video summarization:static video summarization and dynamic one, as well as a comparison between the two summarizations. In addition, in order to further research on keyframe it requires eliminating blurred human face images from keyframe sequence. So it introduces two face blur detection algorithms:DCT (Discrete Cosine Transform) algorithm and edge detection algorithm.Next, the thesis analyzes the video summarization system demand and describes the whole process of video summarization including system design, algorithm verification, system detailed implementation and testing process. In the overall design, the video summarization system is divided into the camara&keyframe management module, video file management module and user interaction module. In the algorithm verification part, it puts forward a keyframe extraction and updating algorithm which is based on partial context and motion detection, and then it puts forward two blur detection algorithms based on DCT and edge detetion aimed at blurred face detection for the keyframe processing. In the detailed design, it describes concrete implementations of the above three modules, mainly including the generation of new summary, the scene boundary definition, optimal summary judgment method and the design of user interfaces covering recording, playback and file Gallery. Then, it describes the testing process of video summarization system, and verifies the usability and user experience of the system.Finally, it summarizes the whole thesis and states the deficiencies of these two systems and the further work for improvement. Meanwhile, it summarizes all the author’s work and academic achievements during the master period.