由于视频图像的知识冗余度,视频压缩可以采用基于图像内容的压缩方式。论文描述了基于内容的压缩系统的整体框架,并且对于视频对象的形状信息,运动信息和纹理信息的编码过程都进行了详细讨论,然后探讨了视频对象的Sprite编码和时间空间可分级编码,最后介绍了针对任意形状视频对象的形状自适应(SA)DCT变换编码,同时指出SA-DCT存在的缺陷并以此作为后继工作的基础。 Due to the knowledge of the video image redundancy, video compression can be based on image content compression. The paper describes the whole framework of content-based compression system, and discusses the encoding process of shape information, motion information and texture information of video object in detail. Then it discusses Sprite coding and time-space scalable coding of video object, and finally The shape adaptive (SA) DCT transform encoding for video objects of arbitrary shape is introduced. At the same time, the defects of SA-DCT are pointed out and used as the basis of subsequent work.