Image Segmentation Based On FUZZY GLSC Histogram With Dynamic Similarity Discrimination Factor

Abstract- Image pressing applications performes image segmentation as pre-processing technique to extract the features for next stage. The application performance depends on image segmentation to process the foreground or background objects. The image segmentation plays a vital role in computer vision and image processing applications. Inspite of having many thresholding techniques in literature they have their own limitations. This paper proposes a new method of thresholding using Gray Level Spatial Correlation GLSC histogram with a dynamic similarity discrimination factor and Fuzzy logic in deciding the threshold using Shannons entropy. The similarity discrimination factor is made dynamic by considering the absolute difference between the global and local mean of the image. Calculating the threshold in the fuzzyfied region makes the segmentation process the most time efficient than the existing methods. Experimental results proove better efficiency than the existing methods. The technique out performs in case of low contrast images.