SOM-based Approach Research Articles

Efficient video compression codebooks using SOM-based vector quantisation

A new rate-constrained self-organising map (SOM) learning algorithm, incorporating a noise-mixing model, is presented as a vector quantiser for very low bit-rate video codecs. A SOM-based approach will exhibit a higher resilience against local minima under low resolution conditions. Practical implementation details and results are also described.

Graph multidimensional scaling with self-organizing maps

Self-organizing maps (SOM) are unsupervised, competitive neural networks used to project high-dimensional data onto a low-dimensional space. In this paper it is shown that SOM can be used to perform multidimensional scaling (MDS) on graphs. The SOM-based approach is applied to two families of random graphs and three real-world networks.

Self-organizing maps for drawing large graphs

Self-organizing maps (SOM) are unsupervised, competitive neural networks used to project high-dimensional data onto a low-dimensional space. In this article we show how SOM can be used to draw graphs in the plane. The SOM-based approach to graph drawing, which belongs to the general class of force-directed algorithms, allows the drawing of arbitrary weighted graphs. It is particularly efficient to draw large graphs and can be used as a preprocessing step before application of a more sophisticated method.