Video processing in codec, which is better than H.264

Abstract

The article is devoted to the task of video compression for surveillance system with stationery cameras by means of effective decreasing of temporal redundancy. The conception of codec is a combination of layered representation and differential pulse-code modulation (DPCM). The layers, from which each frame of decoded video is composed, are alpha channel, background estimation, image with difference between intensities of objects and background estimation, correcting image. Some layers are compressed with lossy version of WebP and the others are compressed with lossless version of WebP. The results of test on real videos shows that proposed codec outperforms MPEG4 and H.264 in the desired range of values of compression ratio.