Study On Video Shot Boundary Detection Algorithm Based On Color Feature Analysis

In the multimedia age,video processing technology has developed to be an important research field of computer science.Video shot boundary detection is an important means to extract video structured information,and the accuracy of the shot boundary detection has a great effect on the follow-up video processing.Usually,shot boundary can be departed into two kinds: abrupt transition and gradual transition.The gradual transition generally refers to the phenomena of fading in,fading out,dissolving and sweeping.Compared with the abrupt transition,the gradual transition has a long duration and a slow change,so that the detection is more difficult.In addition,the video content is always influenced by a variety of external factors,such as noise,text inserts,mark inserts,etc.,which increasing the difficulty for the shot boundary detection.The main goal of this thesis is to research and design the effective shot boundary detection algorithm to improve the detection performance of the abrupt transition and gradual transition,and to provide the technology for the analysis and processing of video images.The main research contents and contributions of this thesis include following two parts:(1)A video shot boundary detection algorithm is proposed based on fuzzy color distribution chart(FCDC).At first,the color feature of the video frame is extracted by fuzzy color distribution chart.Then,by analyzing the relationship between the different frames,we extract the abrupt transition and gradual transition.It can be seen from the experimental results that this method has a better performance compared with some existed methods.(2)A video tracking based shot boundary detection algorithm is proposed.Through the analysis of the above proposed algorithm and its experimental results,it is found that the algorithm can be further improved.Therefore,with the above proposed method,we propose an algorithm based on video tracking.The algorithm combines the image color feature with the video tracking to carry out the video shot boundary detection.The proposed algorithm can effectively reduce the influence of the video object and the rapid movement of the camera for the gradual detection.At the same time,the binary detection method is introduced to improve the efficiency of the shot boundary detection.In addition,we have considered the situation that there simultaneously exist the color video sequence and the gray video sequence in the video,and propose the differentiated threshold method,which effectively extends the applicability of the algorithm.According to the experimental results,we can see that although the algorithm increases the amount of video tracking,it improves the coverage rate of the detection.