Competitive Neural Trees for Vector Quantization

• Authors: Sven Behnke and Nicolaos B. Karayiannis

 
• Neural Network World, vol. 6, num. 3, pp. 263-277, 1996.

 
• Abstract:
This paper presents a self-organizing neural architecture for vector quantization, called Competitive Neural Tree (CNeT). At the node level, the CNeT employs unsupervised competitive learning.
The CNeT performs hierarchical clustering of the feature vectors presented to it as examples, while its growth is controlled by a splitting criterion.
Because of the tree structure, the prototype in the CNeT close to a given example can be determined by searching only a fraction of the tree. This paper introduces different search methods for the CNeT, which are utilized for training and recall.
The efficiency of CNeTs is illustrated by their use in codebook design required for image compression of gray-scale images based on vector quantization.
 
• Full paper: nf96.pdf