Joint segmentation and image interpretation

The problem of image interpretation is formulated in the framework of modul ar integration and multiresolution. The formulation essentially involves th e concept of reductionism and multiresolution, where the image interpretati on task is broken down into simpler subtasks of segmentation and interpreta tion. Moreover, instead of solving the vision task at the finest resolution Omega, we solve the synergetically coupled vision subtasks at coarser reso lutions Omega - xi for Omega greater than or equal to xi > 0 and use the re sults obtained at resolution (Omega - xi) to solve the vision task at resol ution (Omega - xi + 1). We present a solution to the joint segmentation and interpretation problem in the proposed framework. For the interpretation p art we exploit the Markov random held (MRF) based image interpretation sche me developed by Modestino and Zhang [IEEE Trans Pattern Anal. Mach. Intell. pp. 606-615 (1992)]. Experimental results on both indoor and outdoor image s are presented to Validate the proposed framework. (C) 1999 Pattern Recogn ition Society. Published by Elsevier Science Ltd. All rights reserved.