Color clustering using self-organizing maps
Xiaoyu Zhang, Jiusheng Chen, Jiankang Dong · 2007
The Self-Organizing Map (SOM) is a powerful tool for exploratory data analysis which has been employed in a wide range of color clustering. SOM, which is an unsupervised neural network mapping a set of n-dimensional vectors to a two-dimensional topographic map, can achieve the near-optimal segmentation with low computational cost. We point out that the number of output units used in a SOM influences its applicability for clustering. By proposing a clustering method that efficiently classifies image objects with an unknown probability distribution, without requiring the determination of complicated parameters, we demonstrate that SOM can be used for clustering. To ensure that this clustering method is efficient and highly reliable, we define a hierarchical SOM and use it to construct the clustering method. The experimental results show that the system has the desired ability for the clustering of color in a variety of vision tasks.