Improved saliency mechanism for computer vision
Emami, Mohsen · SHAREOK (University of Oklahoma; Oklahoma State University; Central Oklahoma University) · 2013
The objective of this project is to find an efficient biologically plausible model for the bottom-up saliency mechanism of the human vision system (HVS) and employ it in computer vision applications. In practice, analyzing or storing all information entering the human eye at every moment is beyond the capabilities of the human neural system. The saliency mechanism controls the process of selecting and allocating attention to the most "prominent" locations in the scene, which are mostly referred to as "salient points" or "interesting points" in the literature. The same problem of information overload exists in most of the computer vision applications and an efficient visual saliency model can help reducing time consumption of the algorithm. These applications comprise, but are not limited to, automatic target detection, robotics and image and video compression.