Predictive vector quantization using a neural network

  • Authors:

  • Nader Mohsenian;Nasser M. Nasrabadi

  • Affiliations:

  • Department of Electrical Engineering, Princeton University, Princeton, NJ;Department of Electrical and Computer Engineering, State University of New York at Buffalo, Amherst, NY

  • Venue:
  • ICASSP'93 Proceedings of the 1993 IEEE international conference on Acoustics, speech, and signal processing: image and multidimensional signal processing - Volume V
  • Year:
  • 1993

Quantified Score

Hi-index 0.00

Visualization

Abstract

We consider predictive vector quantization (PVQ) of images using two new coding approaches. The first scheme, namely, address-PVQ, exploits the inter-vector (block) dependencies via predicting the VQ address of the current block from the addresses of the previoualy encoded block. A three-layer perceptron WM used M an address-predictor with the position of the residual address being encoded. The second scheme is a vector extension of a DPCM system. It exploits the intervector dependencies via predicting the current block of pixels. The predictive phaae utilizes a three-layer perceptron while the residual block. were vector huantized using the Kohonen Self-Organizing Feature Maps (KSOFM) clusteringalgorithm. The joint-optimization problem for design of the two components of PVQ Was also conaidered. Coding results are presented for monochrome images.