Abstract: Image retrieval is a way of recognizing pre-defined template patterns in the reference images. This technique has been developed rapidly over the past few decades and its applications have been extended to the fields such as face recognition, pulmonary nodules detection, handwriting identification and road detection. Image retrieval is a fundamental issue in pattern recognition. In this work, lateral inhibition (LI) model is adopted as a pre-processing step, which widens the gray level gradients so as to facilitate the subsequent image retrieval scheme. Regarding the search process for perfect geometric transformations that yield a perfect match between the test image and the predefined template, we consider utilizing meta heuristic algorithms, aiming to efficiently promote the probability to obtain global optimums. The artificial bee colony (ABC) algorithm is a bio-inspired optimization technique, which assimilates the foraging behavior of honey bee swarms. It is well known that the algorithm is good at exploration but poor at exploitation. ABC is a swarm intelligence algorithm inspired by the forging behavior of bees. In this algorithm, the employed bees, the onlooker bees, and the scout bees cooperate to search for the optimal nectar source in the space. We present a balance-evolution artificial bee colony (BE-ABC) algorithm that aims to strike a balance between exploration and exploitation rather than just focusing on improving the latter. This BE-ABC algorithm utilizes convergence information during the optimization process to Manipulate its search intensity in the exploration and exploitation phases. Besides that, it incorporates an overall degradation procedure for generating scout bees so as to efficiently prevent premature convergence. Experimental results confirm that BE-ABC algorithm is more capable than several state-of-the-art intelligent algorithms in this LI-based image retrieval scheme. Besides, investigations are also made on the advantages and limitations of this LI model.

Keywords: pulmonary nodules detection, handwriting identification and road detection, balance-evolution artificial bee colony (BE-ABC) algorithm.