Avs Decoder Optimization On The C64 Platform And Text Extraction Algorithm

With the development of mutil-media technology, storage, transport and searches on mutil-media information that contains speech, image, video and text has become the focus of people's attention. Most of mutil-media information is constituted of image and video. This information has a huge data measure and a lot of information concealed in the content of image or video. How to process the information will promote the development of video compressed codeing technology and searches based on the content of image.The development and widely application of video compressed coding technology also promote the establishment of relative standards. AVS is the information compressing standard that our country owns its independence knowledge property right. Part 2 of AVS is the standard of video coding, and it has high compressibility and low complexity, so it can be widely used on digital television, network vidicon and other fields. It can be implemented on embedded platform very conveniently. In all embedded platforms, implementing on DSP can flexibly achieve high efficiency. To achieve real-time require, optimization of the codec on DSP is required.As a lot of information is contained in the text embedded in image, how to get the text is the first stage of research on image searched based on content.This paper first introduces the video codec technology and its implement on DSP platform, the problems in text detection arithmetic is also introduced. Following is the optimazation of AVS decoder on C64 DSP platform, containing AVS codec, C64 DSP platform, the transplant to DSP platform and the modification to suit for DSP after transplanting, and then focus on optimazation based on C64. Next is the investigation of text detection. Besides the introduce of current method on text detection, this paper present the way to detect the caption in grey image.