Font Size: a A A

Design And Implementation For Video Summarization Based On Key Frame On Mobile Device

Posted on:2014-07-24Degree:MasterType:Thesis
Country:ChinaCandidate:L FuFull Text:PDF
GTID:2268330401965739Subject:Optical Engineering
Abstract/Summary:PDF Full Text Request
In recent years, with the development of video summarization technology andmobile Internet technology, it’s possible to realize the video abstract technology inmobile devices.Video summarization on traditional PC platform is lack of userinteraction, while mobile device can compensate for this deficiency. It can also solve theproblem that there are certain restrictions on the mobile equipment storage resource andnetwork traffic. Meanwhile, the technology can be combined with mobile phonemultimedia and bluetooth technology to generate more applications. Videosummarization on mobile platform has its own difficulty, such as moblie device haslimited resources and its storage and computing power are lagging behind PC.According to the characteristics of mobile platform, we proposed a new division modefor video summarization (online mode and offline mode) and achive it on mobile device.Online mode is used for personal video, keyframes are extracted in real time when userare shooting. Offline mode extracts key frames based on patially decoding forprofessional video stored on mobile devices. The main research work on this paper is asfollows:1. Based on the conversion of user’s interest is often accompanied by cameramovement when users shooting, we use diamond search algorithm and SAD matchingcriterion calculating the motion vector of the adjacent images in the video sequence todetermine the carema status during the process of filming and combine imagedefference to achieve a real-time video segmentation algorithm.2. We implement a keyframe dynamic updating algorithm, using clarity andentropy evaluate image quality objectively to select better keyframe to replace theprevious key frame in each sun-section. User can manually add or delete keyframes.3. We present a new algorithm to extract key frame quickly base on partialdecoding for those video stored on mobile device. Algorithm seclects part of I frameaccording to the length of the video as candidate frame, then use the content differentprinciples to select key frames.4. The online mode and offline mode engine code is encapsulated into the the DLL by COM for UI. We build a video summary system on windows phone7.5plaform andmake video player software with keyframe to facilitate the management the video.
Keywords/Search Tags:video summarization, motion vector, mobile device, color historam
PDF Full Text Request
Related items