Font Size: a A A

Research On The Theory And Technologies Of Digital Video Compression

Posted on:2003-05-14Degree:DoctorType:Dissertation
Country:ChinaCandidate:H W ZhaoFull Text:PDF
GTID:1118360095452308Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Digital video coding is the key technology of all kinds of digital video application. In the past twenty years, digital video compression technology has made great progress, and has accelerated the research and development of videophone, videoconference, digital TV, VCD/DVD, and etc. The MPEG standards established by ISO/IEC and ITU have made great influence on digital video application. In the first part of this paper, the status of this technical field was researched and several private opinions were presented.Object-based video coding techniques used in MPEG-4 standard represent the direction of video coding techniques. The video objects in MPEG-4 may be the objects or background in video, so there are three kinds of data about video objects need to be coded: motion, texture, and shape. Shape coding techniques are thoroughly researched in this paper. A new transform called "Zh-transform" which can operate on binary bitmap is presented and a new shape coding method is designed and implemented based on Zh-transform.Like other transforms such as DCT and WT, Zh-transform can transform the image data from spatial domain to another frequency-like domain, and Zh-transformed data can describe the shape from coarser to finer. This is very important for the next coding step. Another advantage of Zh-transform is that Zh-transform has very low computational complexity.In Zh-transform based shape coding method presented in this paper, the shape of video object is represented by a binary bitmap like other bitmap-based methods, but the shortage in scalability of traditional bitmap-based methods has been overcomed. The output stream of Zh-transform based method can be easily divided into one base layer and several enhancement layers according to requirements. Another advantage of Zh-transform based method is that it has higher coding efficiency than the method used in MPEG-4 first edition. Zh-transformed data are reorganised into binary bitmaps and coded with CAE encoder, now CAE encoder can predict the next bit more precisely according to the bits have been coded, so coding efficiency of CAE encoder is improved.Some practical problems in software implementation of Zh-transform based shape coding method and some effort to improve the method are also discussed in this paper.
Keywords/Search Tags:Digital video compression, shape coding, scalability, digital television
PDF Full Text Request
Related items