基于小波域的视频关键帧选择算法

Selecting Key Frame Algorithm in Video Based on Wavelet

视频关键帧的选择算法基于小波变换(discrete wavelet transform,DWT)特征、多分辨率处理及统计技术。算法在不同分辨率下结合视觉掩蔽特性、视频帧相关性分析小波系数并统计,根据统计结果选择合适的阈值,最后通过比较选择出视频关键帧。结果表明,该算法能选择视频关健帧,选择效果优于以像素估计、块估计等与图像内容无关的算法,且简单易行。

This paper presented a method for selecting key frame in video based on multiple scales process,discrete wavelet transform（DWT） and statistical technology.The algorithm combined visual masking and video frame correlation to analysis wavelet coefficient and select proper threshold to deside key frame in video by comparison.Experimental results show that algorithm can select key frame in video and perform better than that based on gradient method or block estimation algorithm.