| Video compression bases on two factors. One is the spatial redundancy of every image; the other is the temporal redundancy of neighbor frame images in video sequences. The aim of video compression coding is to reduce the bit rate of video sequences. So that video can be real time transferred in fixed communication channel.The core of H.263 is motion estimation algorithm and DCT transform coding. When an image is taken by DCT transform, the stronger correlation among the low frequency coefficients is not eliminated. Considering the similar between DCT and DWT, an embedded coding method based on the similar wavelet coefficients structure is proposed. The method rearranges coefficients of 8×8 block DCT and reduces high correlation between low frequency coefficients by a sum of absolute difference method. Then, image is compressed by SPIHT coding with high PSNR at high compression ratios. Experiment results show significantly improved performance in image quality.According to the temporal-spatial correlation between the neighbor images of video sequences and the macro-blocks of an image and the most of search algorithm getting into local smallness, a new motion estimation algorithm based on temporal-spatial correlation property is proposed. Each frame image, which is taken from standard image sequences, is separated by block. This paper classifies the macro-blocks by using temporal-spatial correlation,and predicts the initial point. It selects different search modes... |
PDF Full Text Request