Saliency detection using contrast enhancement and texture smoothing operations

Abstract

An effective framework to improve the saliency maps in complex (low-contrast, small object, and similar background etc.) images is proposed. In proposed scheme, firstly weighted approximated histogram equalization is used to enhance image contrast. Secondly, edge-preserving guided filter is used to minimize the unwanted details (texture) while maintaining the edges and semantics. Afterward, iterative rolling guidance filter is also applied to perform scale-aware local operations for image abstraction. Cellular automata is then used to obtain and optimize saliency cues by exploiting local similarity. Visual and quantitative analysis with state-of-the-art existing techniques shows the significance of proposed technique.