Neuron splitting for efficient feature map formation

Kohonen's Self Organising Feature Map (SOFM) produces an ordered mapping from one space to another. The paper describes an algorithm inspired by the splitting initialisation for the classical LBG method or generalised Lloyd algorithm (Y. Linde et al., 1980) for vector quantiser design, which allows the efficient generation of maps with various topologies and with high local and global ordering.<<ETX>>

[1]  G. S. Stiles,et al.  Fast full search equivalent encoding algorithms for image compression using vector quantization , 1992, IEEE Trans. Image Process..

[2]  K. K. Truong.  Multilayer Kohonen image codebooks with a logarithmic search complexity , 1991, [Proceedings] ICASSP 91: 1991 International Conference on Acoustics, Speech, and Signal Processing.

[3]  S. P. Lloyd,et al.  Least squares quantization in PCM , 1982, IEEE Trans. Inf. Theory.

[4]  R. Gray,et al.  Speech coding based upon vector quantization , 1980, ICASSP.

[5]  D. S. Bradburn.  Reducing transmission error effects using a self-organizing network , 1989, International 1989 Joint Conference on Neural Networks.

[6]  R. W. Harris,et al.  A comparison of several vector quantization codebook generation approaches , 1993, IEEE Trans. Image Process..

[7]  J. V. Black.  Comparison of the performance of vector quantiser training algorithms , 1993 .

[8]  Robert M. Gray,et al.  An Algorithm for Vector Quantizer Design , 1980, IEEE Trans. Commun..