Retracted: Object Detection in Real Time Images using Saliency Mapping Technique

As indicated by the World Health Organization (WHO), 285 million individuals are evaluated to have vision problems and among these 39 million are visually impaired and 246 million have low vision. Visual saliency is the ability of a vision system (human or machine) to select a certain subset of visual information. Whereas an accurate localization of the salient object from an image is a difficult a problem, when the saliency map is noisy and incomplete. The objective of this research is to build up a framework, which utilizes constant recordings to recognize the articles around and direct the visually impaired individuals to explore and evade obstacle impact. Here an effective algorithm is proposed, which detects the objects and obstacles in salient regions by decomposing the input video into the background and residual videos in much lesser time without sacrificing the accuracy of the decomposition. Here after extracting the feature vectors the corresponding saliency map is constructed. A lot of saliency map is processed at various scales for each element and the best saliency map for each element channel are intensity channels, color channels and motion channels. Further, the classification of saliency map is undertaken, wherein objects detected are grouped accordingly. Finally, the optimal visual saliency of a corresponding visual scene is to calculated, which helps the user to understand the characteristics of the respective object. This is a literature review paper of proposed research.