Research And Implementation On Video Coding Part Of The Integrated Image Compression System

As two different types of image data, still images and video have been studied and compressed in independent way according to the currently published international standards. However, in some cases, the two types of images need to be compressed in single system because of the variability of input image sources. If using two separate compression structures for two types of images in one system, it will, on one hand, increase the complexity, cost, size, and power consumption; on the other hand, the traditional methods of video compression is difficult to meet quality requirements. That means a separate design framework can not work well for the special situations. Based on deep research on still images and video compressions algorithms, an integrated compression system is proposed in this paper, which can work in an adaptive mode for still images and video compression. This scheme uses one single compression structure for two different types of images, achieving high efficiency and high performance. Also it has low complexity, simple algorithm, and easy implementation in hardware architecture to meet the needs of variable image compression.The work of this paper is mainly to study and implement the video coding algorithms. We pays more attention to motion estimation algorithm, which is the key modules in video compression, and the algorithm migration and optimization on DSP. The experiment results show that the speed of the DSP program, could be increased more than 6 times after the optimization and could meet the real-time requirement for the integrated compression system.