2005 Fiscal Year Final Research Report Summary
A Study on Video Coding Based on Combination of Motion Compensation and Waveform Coding
| Project/Area Number |
14550376
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | Tokyo University of Science |
|---|
Principal Investigator |
MATSUDA Ichiro Tokyo University of Science, Faculty of Science and Technology, Lecturer, 理工学部, 講師 (70287473)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ITOH Susumu Tokyo University of Science, Faculty of Science and Technology, Professor, 理工学部, 教授 (00147513)
|
|---|
| Project Period (FY) |
2002 – 2005
|
| Keywords | Video coding / Motion-compensated prediction / Iterative motion compensation / Waveform coding / Matching pursuits / Rate control / Arithmetic code / Rate-distortion optimization |
|---|
| Research Abstract |
In this study, we have proposed a new video coding scheme based on seamless combination of motion compensation and waveform coding. The obtained results are summarized as follows.
1.Investigation of the matching pursuits algorithm as a waveform coding method
We investigated an algorithm called matching pursuits, which iteratively approximates a signal using overcomplete basis-functions, and modified it for the proposed scheme.
2.Development of an Iterative motion compensation technique
An iterative motion-compensated prediction method was developed. In this scheme, a decoded image of the previous frame is used as an initial prediction of the current frame and the prediction errors are iteratively reduced by applying block-based motion compensation of local areas.
3.Combination of motion compensation and waveform coding
We combined the above two techniques for efficient video coding. In order to perform one of the two techniques alternatively at each iteration step of the proposed coding scheme, a cost-function which represents the rate-distortion slope was introduced. It enables well-balanced bit allocation between motion compensation and waveform coding.
4.Improvement of coding efficiency
To improve coding efficiency, several techniques such as matching pursuits using non-separable basis-functions, overlapped block motion compensation, motion estimation with sub-pel accuracy and adaptive entropy coder using arithmetic code were introduced into the proposed scheme.
5.Performance evaluation using the software-based codec
The proposed video coding scheme was finally implemented as software-based codec and tested for various video sequences. As a result, it was shown that the proposed scheme outperforms the H.264/AVC scheme which is known as the most efficient video coding standard in terms of a rate-distortion sense.
|
