An Efficient Shape-Based Approach to Image Retrieval

 

Abstract

We consider the problem of finding the best match for a given query shape among candidate shapes stored in a shape base. This is central to a wide range of applications, such as, digital libraries, digital film databases, environmental sciences, and satellite image repositories. We present an efficient matching algorithm built around a novel similarity criterion and based on shape normalization about the shape's diameter, which reduces the effects of noise or limited accuracy during the shape extraction procedure. Our matching algorithm works by gradually “fattening” the query shape until the best match is discovered. The algorithm exhibits poly-logarithmic time behavior assuming uniform distribution of the shape vertices in the locus of their normalized positions.

Author Keywords: Image retrieval; Shape-based matching; Average Hausdorff; Polylog algorithm; Geometric hashing

Article Outline

1. Introduction 2. Similarity criteria 3. Efficient retrieval of similar shapes

3.1. Populating the shape database 3.2. Outline of the matching algorithm 3.3. Computing the initial width 2ε_s of the ε-envelope 3.4. Processing the shapes 3.5. Increasing ε when there are no candidate shapes 3.6. Geometric hashing 3.7. Time complexity of the matching algorithm

4. A prototype system for shape-based image retrieval

4.1. User dialogue 4.2. Experiments

5. Conclusions and future work References