Human-Like Selective Attention Model with Reinforcement and Inhibition Mechanism

In this paper, we propose a trainable selective attention model that can not only inhibit an unwanted salient area but also reinforce an interesting area. The proposed model was implemented by the bottom-up saliency map model in conjunction with the top-down attention mechanism. The bottom-up saliency map model generates a salient area, and human supervisor decides whether the selected salient area is inhibited or reinforced. The fuzzy adaptive resonance theory (Fuzzy-ART) network can generate an inhibit signal or a reinforcement signal so that the sequence of attention areas is modified to be a desired scan path. Computer simulation results show that the proposed model successfully generates the plausible scan path of salient region.