The Cat's Eye Effect Target Recognition Method Based on Visual Attention

摘要

The Cat's eye effect target recognition method based on visual attention (CTRVA) is proposed. The difference image can be processed by a designed second-directional derivative filter at eight directional channels. Morphological method is employed to deal with the filtered image in all directions, which ensures that target can be easily distinguished from background. The salient maps for each channel where the potential targets exist are calculated through the spectral residual approach, and the "target-saliency" map is computed by a designed saliency fusing method. The coarse detection is performed by the adaptive threshold to extract candidate targets from the "target-saliency" map. The real target region is identified by the characteristics of the cat's eye effect target. Experimental results show that the proposed method is efficient and has an outstanding performance for cat's eye effect target detection.