论文部分内容阅读
当前主流的视频压缩标准,如H.264、HEVC等,主要是针对摄像机拍摄的视频而设计的.而屏幕内容(SC)的视频是由计算机生成的,无噪声,要求更高的压缩效率,同时与摄像机拍摄的视频相比包含更多的结构相似性.提出了一种利用屏幕视频中相似性的模板匹配帧内编码方法,采用由21个像素点组成的模板,以像素点为最小单位对整帧进行模板匹配,并使用哈希函数实现高效的模板查找,最后对残差图像进行熵编码.实验结果显示,与标准的HEVC扩展方法相比,该方法在无损编码上压缩率可以提高一倍,编码时间减小了近一半.
Current mainstream video compression standards, such as H.264 and HEVC, are designed primarily for video shot by video cameras, while screen content (SC) videos are computer generated, noiseless and require higher compression efficiency, And contains more structural similarities compared with the video shot by the camera.A template matching intra-frame coding method based on similarity in the screen video is proposed, using a template made up of 21 pixels, with pixel as a minimum unit The template matching is done on the whole frame and a hash function is used to realize efficient template search.Finally, the residual image is entropy coded.Experimental results show that compared with the standard HEVC extension method, this method can improve the compression ratio in lossless coding Double the encoding time by nearly half.