ABSTRACT:

Understanding the subjective meaning of a visual question, by converting it into numerical parameters which will be extracted and compared by a computer, is that the paramount challenge in the sector of intelligent Image Retrieval, conjointly called the ¿semantic gap¿ problem. In this paper, an innovative approach is proposed that mixes a relevance feedback (RF) approach with an evolutionary stochastic algorithm, called particle swarm optimizer (PSO), as a manner to know user's semantics through optimized iterative learning. The retrieval uses human interaction to realize a twofold goal: 1) to guide the swarm particles within the exploration of the solution house towards the cluster of relevant pictures; 2) to dynamically modify the feature house by appropriately weighting the descriptive options according to the users' perception of relevance. Extensive simulations showed that the proposed technique outperforms ancient deterministic RF approaches of the same category, because of its stochastic nature, which permits a higher exploration of complex, nonlinear, and highly-dimensional resolution areas.