Graph Cuts Method for Unsupervised Multiphase Segmentation
Lydia Liz Lukose, Niya Joseph, Kethsy Prabavathy · CiiT international journal of digital image processing · 2011
Here we introduce a novel approach for multiphase image segmentation. The approach presents a unified framework that unifies three basic segmentation approaches; level set methods, graph cut algorithms and unsupervised multiphase segmentation. Here we have presented an image segmentation approach that have the advantages of segmenting the image into multiple phases where the favorable number of segments the image to be segmented will be automatically chooses by the algorithm itself. The unsupervised multiphase segmentation will provide the separation of image into favorable number of multiple phases. Our main objective is to segment the image robust to noise, blurred edges and topology changes and to achieve global optimization and speed.. Here we are segmenting the image by constructing graph on the basis of the intensity values of each pixel in the image. The graph cuts method will provide robustness towards noise and the brute force numerical algorithm for unsupervised multiphase segmentation will provide fast and efficient segmentation and global optimization