A saliency map method with cortex-like mechanisms and sparse representation

Abstract

In the field of visual attention, bottom-up or saliency-based visual attention allows primates to detect non-specific conspicuous objects or targets in cluttered scenes. Simple multi-scale maps detect local spatial discontinuities in intensity, color, orientation, and are combined into a map. HMAX is a feature extraction method and this method is motivated by a quantitative model of visual cortex. In this paper, we introduce the Saliency Criteria to measure the perspective fields. This model is based on cortex-like mechanisms and sparse representation, Saliency Criteria is obtained from Shannon's self-information and entropy. We demonstrate that the proposed model achieves superior accuracy with the comparison to classical approach in static saliency map generation on real data of natural scenes and psychology stimuli patterns.