Quadtree-Based Inexact Graph Matching for Image Analysis

Abstract

This paper presents a new method for segmentation and recognition of image objects based on structural pattern recognition. The input image is decomposed into regions through a quadtree algorithm. The decomposed image is represented by an attributed relational graph (ARG) named input graph. The objects to be recognized are also stored in an ARG named model graph. Object segmentation and recognition are accomplished by matching the input graph to the model graph. The possible inexact matches between the two graphs are cliques of the association graph between them. An objective function, to be optimized, is defined for each clique in order to measure how suitable is the match between the graphs. Therefore, recognition is modeled as an optimization procedure. A beam-search algorithm is used to optimize the objective function. Experimental results corroborating the proposed approach are presented.