Research On A Video Fingerprint Extraction Technology Based On Energy Minimization Segmentation Using Frame Difference Elimination Background Extraction Aligorithm

User generated content is an important information source for the WWW. Interaction is greatly enhanced by distributed uploading.All sorts of information are provided for various needs. Gigantic digital video site is an example of this phenomena.Sites like Youtube and Youku are loved by users all around the world and have piled up tremendous amount of video clips..Easy uploading and loose content control have led to the negative effect of copyright crisis. Copyright protected videos are duplicated and spread across the internet for free, causing huge damage to copyright owners. To compensate this situation, an efficient mechanism is required to weed out the pirate video contents.This mechanism is usually referred to as video feature extraction. Hence its sole purpose is to establish a unique describing agent for each video file, this process is also known as video fingerprint extraction. There has been significant preceeding studies on this topic. Trough the summary of existing technologies, it is found that early algorithms focus either on single frame freature extraction or compareing videos in mathematics domain. As a result, these methods are burdened with high computation cost and exceeding amount of redundant information. To solve this problem, the video segmentation technology is introduced to speed up feature extraction process.Ultimate goal of this technology is full synchronization between automated segmentation and semantic units, realizing robust yet highly effcient video fingerprint extraction.This paper introduces a new segmentation algorithm and corresponding video fingerprint structure based on Energy Minimization Segmentation(EMS) method, which combined edge detection and color domain information to ajust the video segmentation result to mimic the result of human inspenction, in order to satisfy the needs of pirate video contents detection.Based on matlab simulation, this paper augmented EMS technology through theoretical deduction and simulation tests, in order to upgrade its efficiency while maintaining robustness against video post processing.