Saliency Based Human Detection Using The Guided Search System

Since the necessity to differentiate between objects and human becomes prominent in fields like object tracking, CCTV, and pedestrian detection. This work aims to implement a method for detecting human in static medium. In this paper, a bottom up approach has been adopted for detecting salient regions which are the most prominent regions in an image. This approach uses dominancy feature properties of colour and orientation to detect the salient regions. The guided search system which makes use of the top down approach is employed on salient regions to assure human presence. The top down approach has a learning phase in which weights are calculated for images with human presence and a detection phase that’s assures human presence for provided salient regions.