Conference Paper

S. C. Ahalt and J. E. Fowler, "Vector Quantization using Artificial Neural Network Models," in Proceedings of the International Workshop on Adaptive Methods and Emergent Techniques for Signal Processing and Communications, D. Docampo and A. R. Figueras, Eds., Bayona, Spain, June 1993, pp. 42-61.

• Abstract:
  This paper focuses on recent developments in the use of Artificial Neural Networks (ANNs) for Vector Quantization (VQ). A review of the fundamental ANN models used for VQ is presented, including Competitive Learning networks, Kohonen Self-Organizing Feature Maps, and Conscience Techniques including the FSCL algorithm. The paper also briefly reviews the use of VQ-based clustering techniques in classifiers, including Learning Vector Quantizers, Radial Basis Function Classifiers, and the ART architectures.

  The paper then addresses some of the difficulties associated with the use of vector quantization in practical applications. In particular we focus on the use of VQ techniques for image data compression. While it has long been argued that one of the attractive features of ANNs is that they are readily adaptable to real-time hardware, this goal has been elusive. This paper discusses our successful efforts to construct a real-time Differential Vector Quantizer (DVQ) using ANN-inspired VLSI processors. This discussion concentrates on coding performance and entropy characteristics, and we also describe the DVQ architecture we have implemented.

• Text:

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.