TUMOR DETECTION IN MEDICAL IMAGES USING SELF ORGANIZING MAP

Abstract

Boundary is very commonly defined as line that distinguishes two different regions. Boundary provides clarity to human
eye in understanding any view. It plays a major role in Medical field, as finding the correct boundary in noisy images is still a
difficult task. This paper introduces the new technique of detection using the information of intensity and texture of an image.
Our proposed technique detects the boundaries of objects in noisy images using the information from the intensity gradient via
the vector image model and the texture gradient via the edge map. we discuss the proposed technique on various medical images
using Self organizing map (SOM) clustering provides correct boundaries even in an ill-defined images and multi grey level
images. This method is robust and applicable on various kinds of noisy images without prior knowledge of noise properties.