Quantitative properties of Kohonen's self-organizing maps as adaptive vector quantizers.
Toshiyuki TanakaMasao SaitoPublished in: Systems and Computers in Japan (1993)
Keyphrases
- self organizing maps
- neural gas
- vector quantizer
- vector quantization
- image coding
- neural network
- entropy constrained
- unsupervised learning
- competitive learning
- growing neural gas
- feature maps
- distortion measure
- image compression
- input data
- variable length
- finite state
- k means
- tree structured vector quantization
- frequency domain
- anomaly detection
- model selection
- maximum likelihood
- pattern recognition
- learning algorithm