Progressive Wavelet Coefficients Codec System Design and Its Hardware Design and Implementation
碩士 === 國立雲林科技大學 === 電子與資訊工程研究所碩士班 === 92 === In this paper, a novel shape-adaptive zero tree coding scheme and its hardware realization are presented for the discrete wavelet transform (DWT) based image compression. The shape adaptive scheme can totally eliminate the coding overheads of pixels outsi...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2004
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/14670550100519572405 |
| Summary: | 碩士 === 國立雲林科技大學 === 電子與資訊工程研究所碩士班 === 92 === In this paper, a novel shape-adaptive zero tree coding scheme and its hardware realization are presented for the discrete wavelet transform (DWT) based image compression. The shape adaptive scheme can totally eliminate the coding overheads of pixels outside of a video object to save bit rate. The proposed scheme employs a bottom up and breadth first scanning order and each coefficient is examined only once. It can avoid the expensive list buffers required in the SPIHT to save the memory complexity in hardware implementation. The proposed coding scheme is also extended to the 3-D case to handle the medical volumetric data such as MRI. Simulation results indicate that under given bit rates, the plain version (with shape adaptive feature disabled) of the proposed scheme performs slightly better than the SPIHT scheme. Additional 4dB PSNR performance edge over the SPHIT algorithm can be obtained by shape adaptive processing. Its hardwired design is also developed and verified by FPGA. A sustained processing rate of 20 1024x1024 sized frames/sec can be achieved by the derived hardwired design.
|
|---|