Publication

Abstract : Segmentation is an essential task in image analysis process. Due to non-homogeneous intensities, blurred boundaries, noise and minimum of contrast it is a challenging task for image analysists. It has wide range of applications in all fields exclusively in the field medical imaging for disease deionization and early detection. The root cause for non-homogeneous intensities is uncertainty. Various tools have been introduced to handle uncertainty. We have introduced type-2 fuzzy based image segmentation process for edge detection in blurred areas of an image. When compared with classical fuzzy set, it has upper and lower membership values. Since it has more membership values it can handle higher level of uncertainty. In this chapter we have proposed equivalence function associated with strong negation relation which will address each intensity of an image through the membership values. This method is verified with thermographic breast cancer image data set and the results were satisfactory.