Contour extraction of images on parallel computer-local, parallel and stochastic algorithm which learns energy parameters

Abstract

The stochastic relaxation algorithm for S. Geman and D. Geman (1984), which is based on the Markov random-field (MRF) model of images, was implemented on a hypercube parallel computer for contour extraction of human face images. The local energy parameters that define the MRF model were estimated by the authors' learning algorithm (1989), while supervised by a desirable object contour as a teaching signal. D. Geman's constrained optimization method (1987) was utilized to avoid flaws in a fast simulated annealing. The contours extracted by the learning stochastic relaxation method were systematically compared with those obtained by several conventional edge detection methods. The contours extracted by the authors' method include few discontinuity points and small amounts of noise, and faithfully represent the true contours. The authors propose an algorithm, multiple-level, multiple-resolution MRF, which is an extension of the original MRF. This model can incorporate a priori knowledge about the global structures in images and still be implemented in a local and parallel mode. >