|
Title:
|
EFFICIENT DECODING OF VARIABLE-LENGTH ENCODED IMAGE DATA ON THE NIOS II SOFT-CORE PROCESSOR |
|
Author(s):
|
Peter Mårtensson , Jens Persson , Shang Xue , Bengt Oelmann |
|
ISBN:
|
972-99353-8-6 |
|
Editors:
|
João M. P. Cardoso |
|
Year:
|
2005 |
|
Edition:
|
Single |
|
Keywords:
|
VLC decoding, FPGA, soft-core processors. |
|
Type:
|
Workshop Paper |
|
First Page:
|
119 |
|
Last Page:
|
122 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This paper presents an investigation of acceleration of variable-length decoding of run-length coded image data in the Nios II embedded processor for Altera FPGA implementation. In the first approach we examine the possibility of accelerating variable-length decoding for the JPEG standard by implementing the code table look-up as a customized instruction. In the second approach we replace the VLC code table with a coding technique, we previously have proposed called Alternating Coding (ALT) and implement another set of customized instructions for accelerating the decoding. The VLC decoder is implemented in software using the C programming language with function calls accessing the hardware accelerated instructions. For decoding using standard JPEG VLC codes, one customized instruction resulted in a speed-up of 2.2 times for decoding the codewords compared to an all software implementation. For the Alternating CodingVLC, the software and hardware accelerated implementations both resulted in a speed-up of more 2.6 times compared to the standard JPEG software implementation. For the ALT coding we conclude that VLC decoding in software can be made as efficient as the hardware accelerated standard JPEG implementation. |
|
|
|
|
|
|
