A New Saliency Detection Model in Remote Sensing Images with Sea Background

Human visual system is very efficient and selective in scene analysis, which has been widely used in image processing. This paper try to combines the characteristics of human visual system and propose a new bottom-up visual saliency model used for remote sensing saliency detection. This model is based on the premise that locally contrasted and globally rare features are salient. First, the low-level features of luminance and chrominance are directly extracted from the image. Second, a Gabor filter bank is applied on the three color channels to extract medium-level features as image orientation information. A comparison based on a 100 images (with typical ocean background) dataset. The experimental results demonstrate that the proposed method performs well in predicting human fixations and recognizing saliency area.