An Evaluation of Image Matching Algorithms for Region Based Image Retrieval

Content based image retrieval systems (CBIR) have drawn wide attention in recent years due to the exponential growth of multimedia information storage in various fields of life. Given a query image, the goal of CBIR system is to retrieve content related images from databases by employing suitable image matching algorithms. Many algorithms have been proposed in the past for image matching. This paper evaluates the performance of three known image matching algorithmsIntegrated region matching algorithm (IRM), Minimum distance algorithm and Integrated image matching algorithmused in region based image retrieval systems. The performances of the algorithms are evaluated in terms of average precision and time required to retrieve images. It is observed that the minimum distance algorithm is more suitable for CBIR than the other two algorithms. Its performance was found to be competent with IRM algorithm in terms of average precision and outperformed both algorithms in terms of retrieval speed, which is a desirable property for CBIR systems.