DocumentCode
2062360
Title
Improving the performance of the quantized biorthogonal 9/7 wavelet filters
Author
Kotteri, K.A. ; Bell, A.E. ; Carletta, J.E.
Author_Institution
Virginia Tech, Blacksburg, VA, USA
fYear
2002
fDate
13-16 Oct. 2002
Firstpage
245
Lastpage
250
Abstract
The biorthogonal 9/7 wavelet is used for lossy compression in the JPEG2000 image coding standard. A hardware implementation of the standard must consider the accuracy and the efficiency with which the quantized filter coefficients are represented; filter structure is also an important design consideration. A high precision representation ensures compression performance close to the unquantized, infinite precision filter bank, but this comes at the cost of increased hardware resources and processing time. This paper proposes a gain compensation method that significantly improves the performance of the quantized filters without sacrificing efficiency. Our best new method achieves performance close to the unquantized filter case while also realizing a fast, efficient hardware implementation.
Keywords
channel bank filters; image coding; transform coding; wavelet transforms; JPEG2000 image coding standard; filter structure; gain compensation method; hardware implementation; lossy compression; quantized biorthogonal 9/7 wavelet filters; quantized filter coefficients; Costs; Discrete wavelet transforms; Field programmable gate arrays; Filter bank; Finite impulse response filter; Hardware; Image coding; PSNR; Performance gain; Quantization;
fLanguage
English
Publisher
ieee
Conference_Titel
Digital Signal Processing Workshop, 2002 and the 2nd Signal Processing Education Workshop. Proceedings of 2002 IEEE 10th
Print_ISBN
0-7803-8116-5
Type
conf
DOI
10.1109/DSPWS.2002.1231112
Filename
1231112
Link To Document