Constraint saliency based intelligent camera for enhancing viewers attention towards intended face

Abstract Visual Saliency decides the focus of attention towards a region in a scene. When we talk about attending faces in a crowd or set of multiple faces, our focus of attention does not go equal for all the faces. The biasness of human's visual system towards a particular face, occurs due to some dominant features of it, over rest of the faces. So, for the faces which are not intrinsically salient in a scene, there is a requirement of increment of attentiveness. The current study effort to improve the saliency of a face (in a set of multiple faces), which is not significantly salient. It can be achieved by enhancing the contrast of the intended face with its surrounding faces, in terms of visual low-level features like intensity, color, etc. Modification of such feature values will change its potential to attract observer's gaze. But excesses change will destroy the originality of the image. Therefore, the problem of enhancing saliency of a target face is framed as an optimization (maximization) problem under some constraints. This concept can be applied to develop a saliency based intelligent camera having the power of enhancing the attractiveness of a particular face in the crowd and in the taken photograph, the enhanced face may give more attention to the viewers. Experiment has been conducted on grayscale as well as the colour images. Moreover, effect of saliency on faces wearing jewellery, has also been measured.