In this paper image segmentation techniques have been explored which uses super pixel as intermediate step along with fuzzy clustering methods. Superpixel segmentation is the process of partitioning an image into multiple segments called superpixels, which are homogeneous as in pixels inside every portion are comparable concerning certain attributes, for example, shading and surface.
In spite of the fact that superpixelsegmentation as a rule yields over-sectioned results instead of item level fragments, it radically diminishes the quantity of picture primitives with insignificant loss of data and offers a simple approach to separate the probably picture objects with as few portions as could be expected under the circumstances. Likewise, since superpixelsegmentation gives a more characteristic and perceptually significant representation of the info picture, it is more helpful and powerful to concentrate area based visual elements utilizing superpixels.
In order to get better segmentation the FCM use the GLCM features of turbo pixels instead of intensity values of pixels and hence help in decision making to put a particular turbo-pixel into different fcm clusters. In the proposed work, we will explore these techniques in order to get better segmentation of different sections of the input images.
Journal published: Mehra, J., & Neeru, N. (2017). A brief review: Super-pixel based image segmentation methods. International Journal Of Research And Engineering, 4(7), 200-205.