Algorithm Research On Video Compression Based On Wavelet Transform

With the rapid development of the computer and communication technology and the popularization of 3G technology in the near future, multi-information, especially video image is very abundant and plays an important role in our dailylife gradually. The rich video information is benefit to people, but the large number is a disadvantage in fact. So the video image compression is very necessary and the compressing technology attracts many pursuers to research.Wavelet transformation has good character of spatial and frequency and supplys mufti-resolution to signal, which is very important for coding. Stochastic video image can be separated various multilevel subbands which are comparatively stabile and easy to coding. Coding those stabile components is more efficient than that of whole signal. Wavelet transformation is more suitable to expandable bitestream, because of its mufti-resolution decomposition. Wavelet transformation has the flexibility in analyzing the stochastic signal of video image and the ability of adapting the human vision, so it is becoming a powerful tool to image and video image coding.This thesis describes some typical compression algorithm of video image and presents a video image compression method of 3D wavelet transformation with temporal enhancement and the rate control of temporal stability. Comparing to other coding algorithm of 3D wavelet transform, this algorithm based on the human visual system (HVS) performs different ranges quantification to different frequency data of video image, to solve the problem that the large number of data is created when motion is acute. Distributing the coding rate based on the sub-block rate distortion ratio balances the distortion of decompressed temporal signal. Experiment shows the following results: firstly, this algorithm can reduce the complexity without motion estimation and compensation; secondly, this approach using the lifting scheme wavelet transform can save memory space and improve operation speed; thirdly, this method can get good temporal stability and improves the articulation and fluency of video image through controling rate. The simulation experiment shows that it is valid and feasible.