Texture based segmentation of cell images using neural networks and mathematical morphology

In this paper we show a new technique of image segmentation that uses a pre-processing algorithm combined with an unsupervised method of segmentation that uses a competitive neural network. The segmentation is based on the texture present in the image. We tested the algorithm in a system built to classify and count the number of cells in the image. After the segmentation, the system uses mathematical morphology techniques to improve the segmentation process. An online version of the system was implemented, making it easily available for those interested in using it