A performance study of image segmentation techniques

Image based applications such as target tracking, tumor detection, texture extraction requires an efficient image segmentation process. The partitioning of image into various non- overlapping distinct regions refers the image segmentation. Various segmentation techniques like edge, threshold, region, clustering and neural network are involved in the effective image analysis. The efficiency of the segmentation process improved with the help of several algorithms, namely, active contour, level set, Fuzzy clustering and K-means clustering. This paper analyses the performance of algorithms for image segmentation in detail. Intensity and texture based image segmentation is the two levels of the level set method. The combination of both intensity and texture based image segmentation provides better results than the traditional methods. The detailed survey of segmentation techniques provides the requirement of the suitable enhancement method that supports both intensity and texture based segmentation for better results. The comparison between the traditional image segmentation techniques are illustrated.