Applications of low density party check codes to data compression with distortion
Project/Area Number |
14550347
|
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 | Kazunori Hoshino, |
Principal Investigator |
YAMAMOTO Hirosuke The University of Tokyo, Graduate School of Information Science and Technology, Professor, 大学院・情報理工学系研究科, 教授 (30136212)
|
Co-Investigator(Kenkyū-buntansha) |
OGAWA Tomohiro The University of Tokyo, Graduate School of Information Science and Technology, Research Associate, 大学院・情報理工学系研究科, 助手 (00323527)
IWATA Satoru The University of Tokyo, Graduate School of Information Science and Technology, Associate Professor, 大学院・情報理工学系研究科, 助教授 (00263161)
|
Project Period (FY) |
2002 – 2003
|
Project Status |
Completed (Fiscal Year 2003)
|
Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 2003: ¥1,600,000 (Direct Cost: ¥1,600,000)
Fiscal Year 2002: ¥1,700,000 (Direct Cost: ¥1,700,000)
|
Keywords | data compression with distortion / rate-distortion function / low density parity check code / source coding theorem / belief propagation / Mackay's ensemble |
Research Abstract |
A LPDC(low density parity check) code is an error correcting code that was first proposed by Gallager in 1962. Recently, BP(belief propagation) or other repeating decoding algorithms become applicable to the LDPC codes by rapid growth of computer power. In this research, we applied the LODC codes to data compression with distortion, and evaluated the performance. In 2002, we proposed an encoding algorithm to apply the LODC codes to data compression with distortion. Furthermore, we proved that there exits a LODC codes that can attain the theoretical optimal rate, i.e., rate-distortion function. In 2003, we improved the proof and published the results in the IEEE transactions on Information Theory. In the coding theorem, it is shown that if column weight t in the parity check matrix of a LODC codes satisfies t=O(log n) for code length n, the code can attain the rate-distortion function asymptotically. Furthermore, if t=O(n), then it converge to the rate-distortion function exponentially. We also show by simulation that the BP algorithm of LPDC codes for error correcting cannot attain a good performance in the case of rate-distortion coding.
|
Report
(3 results)
Research Products
(3 results)