A Hybrid Visual Attention Method Based on Itti Model

An improved visual attention method based on human vision attention model Itti has been proposed. Firstly, the contour information is added to the Itti model to improve the recognition ability. Then an adaptive hybrid strategy for color and brightness feature maps is proposed, according to human retina´s physiological characteristics under different environments. Besides, by adding motion feature maps, the proposed method also is applicable to the videos. The experimental results show that the hybrid visual attention method improves the original Itti model´s ability and more in line with the human eyes.