Novel Method for Keyframe Extraction using Block Truncation Coding and Mean Square Error

摘要

A video signal is made of up ‘n’ number of frames. Frames per second gives the number of frames per second, hence 32 fps means there are 32 frames in one second. In video processing we have to process each frame one by one, but processing each frame would consume a lot of resources and time. Key frame is a frame in which there is a major change as compared to the previous frame. Hence instead of processing all the frames we only process the key frames. In applications like video content summarization, data storage and surveillance key frames extraction plays a vital role. This paper explains key frame extraction by MSE, sobel edge detection and MSE, BTC, DTTBTC and TSTBTC. Results of the methods are compared and discussed.