Salient Region Detection Based on Spatial Weight Map

Saliency detection is defined to be a key attention mechanism by enabling organisms to focus their limited perceptual and cognitive resources of available sensory data. In short, saliency detection is nothing but detecting the more attracted regions in an image. Like a loud noise in a quite environment saliency is the contras-tic difference between the visually attracted items and their neighborhood. Detecting those attracted areas in the image is termed to be saliency detection. We proposed a novel salient region detection method by integrating four features namely boundary information, frequency weight map, global contrast and color spatial variance. Finally, the saliency map is defined as being the average of the three feature maps added by color spatial variance. Experimental result shows that the proposed method produce better performance compared to the state-of-the-art of methods. The programming and simulation of the processes as well as the analysis of the results were done using MATLAB.