A comparative performance evaluation of segmented image with obstacle for textural coarseness

There are wide varieties of textural properties. Texture coarseness is one of them, which is important feature to process the images. There are various methods to calculate the coarseness of an image but all gives different types of results. This paper compares the all four important textural coarseness techniques. Paper describes Gray Level Cooccurrence Matrix (GLCM), Tamura, Auto Correlation and Fractal textural coarseness techniques to calculate the coarseness and compare all of them on a sample image set. Two samples are shown in last section. Sample images are segmented by homogeneity measurement decision. Segmentation method used in this paper is based on color and texture descriptor of the image and segment the image in various parts, showing there is any obstacle present in the image. Our experiment shows if there is any obstacle in the image and differentiate Tamura method from all of these coarseness methods describing best performance followed by fractal method.