Abstract
This paper presents an image segmentation technique using graph tools for object detection. Graph theoretical systems have many good features among different segmentation schemes. It organizes the image elements into mathematically and structural form, and makes the problem formulation more flexible, and the computation more efficient. In this paper, the work, consists of two stages, in the first stage we apply the pixel-based labeling algorithm to the binary image, the algorithm works similarly to the eight connectivity labeling, but it is broader than it in terms of the search area, where two (horizontal and vertical) thresholds are first defined so that the search is in a block whose height is the vertical threshold and its width (2 * horizontal threshold), in the second stage the output of the first stage is mapping into an undirected weighted graph, in which each vertex represent region rather than pixel. We evaluate the results by comparing it with other method using (RI) parameter. We use 50 image from online source and image taken from Berkeley database.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
