Incorporating random forest trees with particle swarm optimization for automatic image annotation

This paper presents an automatic image annotation approach that integrates the random forest classifier with particle swarm optimization algorithm for classes' scores weighting. The proposed hybrid approach refines the output of multi-class classification that is based on the usage of random forest classifier for automatically labeling images with a number of words. Each input image is segmented using the normalized cuts segmentation algorithm in order to create a descriptor for each segment. Images feature vectors are clustered into K clusters and a random forest classifier is trained for each cluster. Particle swarm optimization algorithm is employed as a search strategy to identify an optimal weighting for classes' scores from random forest classifiers. The proposed approach has been applied on Corel5K benchmark dataset. Experimental results and comparative performance evaluation, for results obtained from the proposed approach and other related researches, demonstrate that the proposed approach outperforms the performance of other approaches, considering annotation accuracy, for the experimented dataset.