The Technology Of Video Key-frame Extraction And Its Application In Public Opinion Analysis

With the rapid development of the computer technology and multimedia technology,video multimedia information has been the public opinion focus carrier,the demand for searching proper public sentiment focus video segment in a great of video footage quickly and conveniently has become increasingly urgent,and also the traditional searching method based on character purely can not meet the requirements. In order to solve these problems,content-based video searching technology had appeared in the 1990s,not only using text,also using vision information such as color, texture,shape and boundary to describe the video.Shot detection and key-frame extraction is two key technologies of video processing.This thesis does a further research for shot detection and key-frame extraction based on current algorithm.1) Video play and video frame capture are achieved by using OpenCV technology;2) This paper discusses the principle of Video Shot Segmentation and main algorithm,and has an analysis for current mainstream algorithm and at the same time has an analysis between two algorithms.3) Based on shot detection,this paper does a research for the principle and main key-frame algorithm of key-frame extraction,raise a simple and effective key-frame extraction algorithm,use Visual Studio 2008 as one developing tool to achieve the extraction algorithm,and also does a availability experiment for key-frame extraction algorithm.4) Based on key-frame extraction technology,this paper raises an effective Automatic Shot Clustering algorithm through analyzing the basic method for Shot Clustering.This algorithm can avoid disadvantage of the tradition shot clustering method such as great mass calculation,low efficiency.By using the key-frame and comparing the key-frame,calculation can be reduced effectively,and also shot clustering efficiency can be improved.