A Review on the Strategies and Techniques of Image Segmentation

Abstract

Segmentation is a method of partitioning an image or picture into different regions which has same attributes like Texture, intensity, gray level etc with the motive to yield object of interest from the background. It is a method in which we included the object belongs to the same category in one class and the objects that belong to other category are added in other class for separating the object and background. There are several image segmentation techniques namely traditional thresholding (Otsu) and clustering segmentation (K-means). By differentiating all these image segmentation techniques we have to find which segmentation technique is better on the characteristics of image segmented. Segmentation is done on built in environment which becomes more demanding. In built in environment, both K-means and Otsu are unsuccessful to yield good standard of segmentation because of varying lightening on the image and complex surrounding.