Unsupervised Learning Technique for Image Segmentation

Image segmentation is a fundamental step in many applications of image processing. Many techniques exist for image segmentation based on different methods as classification-based methods, edge-based methods, region-based methods, and hybrid methods. The principal approach of segmentation is based on thresholding (classification) that is lied to the problem of the thresholds estimation. We assume that the data of image can be modeled by Gamma distribution. In this paper, we will explain a new method that estimates thresholds using the unsupervised learning technique (ISODATA) with Gamma distribution. The ISODATA (Iterative Self-Organizing Data Analysis Technique) method is one of the classification-based methods in image segmentation. It is an unsupervised classification algorithm. The objective of this algorithm is to split a non-homogeneous region into two sub-regions by using statistical parameters of the Gamma distribution of two sub-regions. Experimental results showed good segmentation for artificial and real images.