A fast encoding algorithm for vector quantization based on Hadamard transform

For vector quantization (VQ), it is extremely time-consuming to extract the similar codeword with input vector during the encoding process. In this paper, a novel algorithm based on Hadamard transform (HT) is proposed. Two inequalities derived after performing HT to all input vectors are used to ear...

Full description

Saved in:
Bibliographic Details
Published in:2008 34th Annual Conference of IEEE Industrial Electronics pp. 1817 - 1821
Main Authors: Jiann-Der Lee, Yaw-Hwang Chiou
Format: Conference Proceeding
Language:English
Published: IEEE 01.11.2008
Subjects:
Data compression
Encoding
Error correction
Error correction codes
Filtering
Hadamard transform
Image coding
Image storage
Nearest neighbor searches
Vector quantization
Watermarking
ISBN:9781424417674, 1424417678
ISSN:1553-572X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:For vector quantization (VQ), it is extremely time-consuming to extract the similar codeword with input vector during the encoding process. In this paper, a novel algorithm based on Hadamard transform (HT) is proposed. Two inequalities derived after performing HT to all input vectors are used to early remove impossible codewords in the distortion computations stage. From the experimental results, it is shown that the proposed approach can largely decrease the computation time for achieving VQ coding with the same quality with full search algorithm. More specifically, compared with the DHSS algorithm, the proposed algorithm reduces the computational time by 25% to 61%. Compared with the Panpsilas algorithm, the proposed algorithm reduces the computational time by 60% to 70%. Compared with the Laipsilas algorithm, the proposed algorithm reduces the computational time by 42% to 55%. Compared with the HTPDE algorithm, the proposed algorithm reduces the computational time by 20% to 40%.
ISBN:9781424417674
1424417678
ISSN:1553-572X
DOI:10.1109/IECON.2008.4758231