SALIENT REGION DETECTION BASED ON AVERAGE SALIENCY MAP WITH SPATIAL VARIANCE

Salient region detection is one of a pre-processing method which are useful for several image processing application. Image processing application includes image retrieval, object recognition, face recognition, adaptive compression and image retargeting. Thus salient region detection is detecting the more attracted regions in an image. It can be define as contrast difference between the visually attracted regions and its neighbourhood in an image. In this paper we proposed a novel method by creating a saliency map by average of the three feature maps added by colour spatial variance of an image. Main feature maps are frequency map and spectral map.