Guest session — Image processing: A Nash-game approach for image restoration and segmentation

摘要

Image restoration is a very important task in image processing. It consists in removing or reducing the noise from an observed image. In the variational approach to the image restoration problem an energy composed of two terms is minimized. One term expresses fidelity to the observed (noisy) image and the other forces the recovered image to be sufficiently regular. In this work, we propose a game approach to simultaneously restore and segment noisy images. We define two players, the first one is the restoration, the image intensity is a strategy, and the second one is a segmentation with the contours as a strategy. The two players play a static game with complete information, and we consider the Nash Equilibrium as solution. For the computation of this equilibrium we present an iterative algorithm. The results of numerical experiments show the efficiency of the proposed method.