Applications Of Adaptive Transform In Video Compression

With the rapid development of signal processing, communication technology andmultimedia technology, digital video compression technology has become a more andmore porpular. The new generation of video compression standard H.264with theadvantages of high compression efficiency, low compression rate, good quality ofimage reconstruction and high affinity of network, is the most widely used at present.It is mainly applied in digital broadcast television, video on demand, videoconferencing, etc. In recent years, the wavelet analysis theory has gain greatimprovement and discrete wavelet transform in digital image processing has made abreakthrough. It is possible to adopt wavelet transform technology to improve thecompression efficiency of H.264. Therefore, research on the wavelet analysis theoryand its application in the H.264standard is importantly significant for promoting thevideo compression standard.This paper firstly analyzes the algorithm of predicting code and transformingcode in the H.264，then apply the wavelet analysis theory in video compression. Themain work includes:(1) Analyzes the structure of H.264video compression code system, focusing onthe application of four discrete transform (Haar wavelet transform, discrete cosinetransform, discrete Hartley transform and4-ary Optimal Haar wavelet transform) invideo compression.(2) Based on the above four transformation, corresponding quantization matrixwere trained.(3) Respectively, analyzing discrete cosine matrix and a4-ary optimal Haarwavelet matrix by apply them to transform Gaussian signal. Based on the analysis ofthe generated transformation matrix, we found that when the video bit rate is low, thecompression efficiency which is based on the discrete wavelet transform (DWT) issuperior to the the compression efficiency of H.264basing on discrete cosinetransform (DCT) on the whole. But when at higher video bit rate, H.264is better.(4) According to H.264own characteristics of bit stream, we propose an adaptivevideo compression algorithm which combines discrete wavelet transform and discretecosine transform. Experimental results indicate that the overall compressionefficiency of the algorithm is superior to both H.264and compression algorithmwhich is based on discrete wavelet transform no matter for high-bit-rate video or lowbit rate.