Chinese Journal of Computers   Full Text
Title	A Framework of Video Format Adaptation Based on Motion Description
Authors	SHEN Qiu LI Hou-Qiang WANG Yi HE Jia
Address	(Information Proceesing Center, Department of Electronic Engineering & Information Science, University of Science & Technology of China, Heifei230027)
Year	2009
Issue	No.6(1162—1171)
Abstract &   Background	Abstract Due to the existence of various video coding standards in multimedia applications, it is of great significance to implement video format adaptation among these standards. This paper presents a novel framework for video format adaptation, which is based on motion description and is originated from the authors’ early work. Such motion description comprises abundant motion information, and can be obtained by pre-encoding a video sequence. Motion description can be used not only to fast generate a required bitstream, but also to guide the transcoding between different video formats. As an example, format adaptation between H.264/AVC and MPEG-2 which is on the basis of the proposed framework is illustrated in detail. A number of comparison experiments are performed to show that the proposed solution can greatly reduce the computational complexity meanwhile retain the coding efficiency, which manifestly demonstrates the proposed solution is superior to the conventional transcoding technique. Keywords video adaptation; format transcoding; H.264/AVC video coding standard; MPEG-2 video coding standard Background The work of this paper is supported by National Natural Science Foundation of China (NSFC) Key Program under contract No.60632040, NSFC General Program under contract No.60672161. Currently, there exist various video coding standards in multimedia applications, including ITU-T H.26x, ISO/IEC MPEG-x and national standardization initiatives, for example AVS codec standardized by Audio and Video Coding Standard Workgroup in China. These standards play an important role in their respective fields, but cannot be directly applied to other fields. To adapt the application environment with multiple standards, video signal should be coded into several different formats, which would bring a great computational complexity and consume a large amount of storage capacity. Format transcoding tries to transform a video bit-stream from one format to another in an effective and rapid manner. Generally speaking, conventional method should make trade-off between computational complexity and coding efficiency, meanwhile different algorithms should be designed distinctively for different format transcoding, which bring a great burden to real-time transcoding system. This paper attempts to give a novel solution of realizing format adaptation in a general framework. The basis of the framework is an intermediate motion description of video, which is obtained by pre-encoding the original video signal. As the description can provides abundant motion information of the original video signal, while encoding or transcoding a video signal into one format, the complex process of motion estimation can be substituted by extracting motion information from the description with a simple process. Considering H.264/AVC and MPEG-2 are two widely used video coding format, this paper illustrates format adaptation between these two formats in detail. Furthermore, a mapping algorithm of motion information required for MPEG-2 encoding and/or transcoding is described particularly. Experimental results demonstrate that the proposed solution can support video format adaptation rapidly while maintaining coding efficiency. In the future, more mapping algorithms of motion information should be designed to support more format adaptation. Also, the intermediate description can be enhanced by importing new features of video signal, so that multiple video adaptations rather than format adaptation can be supported, for example bit-rate adaptation, temporal and/or spatial resolution adoption, etc.