Interactive Image Segmentation with Minimum Intensity Variation based Region Merging
Pramod Kumar Sethy, Sajan Kor · International journal of engineering and future technology · 2016
Extracting required information from a pool of data is not a very convoluted task when compared to extracting an object from a digital image. For extracting an object from a digital image, a process called Image Segmentation is used. As a basis to this process of image segmentation with minimum intensity variation based region merging, an initial segmentation is performed. In the initial segmentation process, an image is disintegrated into finite number of non-overlapping regions with definitive characteristics. Various methods have been developed for doing this initial image segmentation. This paper deliberates the idea of image segmentation using lowest intensity variation based region merging with minimum inputs by user using simple brush strokes called markers(object and background markers). In the proposed method, an image is initially segmented using some low level segmentation methods and then region merging process is guided by the help of these input markers on initially segmented image. A merging rule for this proposed method is defined as a two stage iterative merging algorithm that gradually label each unmarked region as either object or background to finally contour the object for extraction. Further it also deliberates the robustness of the proposed method by comparing it with the existing segmentation methods.