Texture invariant image segmentation via particle swarm optimization

![CDATA[Segmentation is a critical element in image processing since it can divide the image into various smaller regions for faster processing in later stages. Colour and shape, as the most distinguishable visual features in describing visual objects, have been adopted by many researchers as important information sources in image segmentation. In this work, an image segmentation approach is proposed that uses Particle Swarm Optimization (PSO) to find the appropriate combination of colour and shape information that can provide a proper volumetric description of image contents. The segmentation process is carried out via the computational friendly Otsu's multi-threshold. A spatial channel is computed and added to the PSO search space to address the issue of loss of spatial information led by the thresholding method.]]