Issue No. 01 - January (1999 vol. 21)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/34.745730
<p><b>Abstract</b>—Robust segmentation of structures from an image is essential for a variety of image analysis problems. However, the conventional methods of region-based segmentation and gradient-based boundary finding are often frustrated by poor image quality. Here we propose a method to integrate the two approaches using game theory in an effort to form a unified approach that is robust to noise and poor initialization. This combines the perceptual notions of complete boundary information using edge data and shape priors with gray-level homogeneity using two computational modules. The novelty of the method is that this is a bidirectional framework, whereby both computational modules improve their results through mutual information sharing. A number of experiments were performed both on synthetic datasets and datasets of real images to evaluate the new approach and it is shown that the integrated method typically performs better than conventional gradient-based boundary finding.</p>
Image segmentation, integration, game theory, boundary finding, region-based segmentation, MRF.
J. S. Duncan and A. Chakraborty, "Game-Theoretic Integration for Image Segmentation," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 21, no. , pp. 12-30, 1999.