An adaptive incremental LBG for vector quantization

Authors:
F. Shen; O. Hasegawa
Affiliations:
Department of Computational Intelligence and System Science, Tokyo Institute of Technology, R2, 4259 Nagatsuta, Midori-ku, Yokohama, 226-8503, Japan;Imaging Science and Engineering Lab., Tokyo Institute of Technology and PRESTO, Japan Science and Technology Agency (JST)
Venue:
Neural Networks
Year:
2006

Citing 6
Cited 10

Robust Clustering with Applications in Computer Vision

IEEE Transactions on Pattern Analysis and Machine Intelligence
The LBG-U Method for Vector Quantization – an Improvement over LBGInspired from Neural Networks

Neural Processing Letters
Multi-Stage Target Recognition Using Modular Vector Quantizers and Multilayer Perceptrons.

CVPR '96 Proceedings of the 1996 Conference on Computer Vision and Pattern Recognition (CVPR '96)
Vector quantization of image subbands: a survey

IEEE Transactions on Image Processing
Bayes risk weighted vector quantization with posterior estimation for image compression and classification

IEEE Transactions on Image Processing
`Neural-gas' network for vector quantization and its application to time-series prediction

IEEE Transactions on Neural Networks

An enhanced self-organizing incremental neural network for online unsupervised learning

Neural Networks
Improved batch fuzzy learning vector quantization for image compression

Information Sciences: an International Journal
Expansive competitive learning for kernel vector quantization

Pattern Recognition Letters
A fast VQ codebook generation algorithm via pattern reduction

Pattern Recognition Letters
A Fast VQ Codebook Generation Algorithm Based on Otsu Histogram Threshold

Fundamenta Informaticae
Hidden Markov models applied to snakes behavior identification

PSIVT'07 Proceedings of the 2nd Pacific Rim conference on Advances in image and video technology
Fuzzy vector quantization for image compression based on competitive agglomeration and a novel codeword migration strategy

Engineering Applications of Artificial Intelligence
A Fast VQ Codebook Generation Algorithm Based on Otsu Histogram Threshold

Fundamenta Informaticae
On the systematic development of fast fuzzy vector quantization for grayscale image compression

Neural Networks
A proposition of adaptive state space partition in reinforcement learning with Voronoi tessellation

Artificial Life and Robotics

Quantified Score

Hi-index	0.00

Visualization

Abstract

This study presents a new vector quantization method that generates codewords incrementally. New codewords are inserted in regions of the input vector space where the distortion error is highest until the desired number of codewords (or a distortion error threshold) is achieved. Adoption of the adaptive distance function greatly increases the proposed method's performance. During the incremental process, a removal-insertion technique is used to fine-tune the codebook to make the proposed method independent of initial conditions. The proposed method works better than some recently published efficient algorithms such as Enhanced LBG (Patane, & Russo, 2001) for traditional tasks: with fixed number of codewords, to find a suitable codebook to minimize distortion error. The proposed method can also be used for new tasks that are insoluble using traditional methods: with fixed distortion error, to minimize the number of codewords and find a suitable codebook. Experiments for some image compression problems indicate that the proposed method works well.