Image segmentation using the level set method by the data clustering algorithm
김연후 · Seoul National University Open Repository (Seoul National University) · 2013
This paper discusses a variational method for image segmentation by using the level set method with data clustering algorithm.The traditional method of image segmentation has been to separate the given image into two regions-inside and outside.But in the multi-phase segmentation problem, the result is considerably different depending on the number of phases which is given a priori by the user.Therefore, choosing the exact phase number is a critical decision in the multi-phase image segmentation.In this study, we investigated the previous image segmentation models.Then, a multi-phase segmentation method using data clustering algorithms on the image histogram is presented.This study analyzed previous algorithms by giving some changes on the algorithm, and determined the phase number by applying new segmentation method and data clustering methods.Finally, some numerical results were shown for various images.