Vector Quantization-Based Automatic Shot Boundary Detection

摘要

A compact video description is required to efficiently browse a multimedia database. A possible solution is the creation of a summary by using a set of frames (key frames) representing the content of each part of the video. To this end, a novel shot boundary detection method, based on a vector quantization approach, is presented. Experimental results show the effectiveness, in terms of computational cost and performance, of the proposed system.