Edge-Based Gaze Planning for Salient Proto-Objects

Abstract

Selective attention of primates or human being has been modeled and implemented by many vision researchers in a decade. And also several attempts, which improve object recognition using selective attention model, have existed. But there are few researches in the gaze planning for the results of selective attention process. Therefore we propose a planning method based on edge information in the attended regions. And we explain the edge description methods (edge density and entropy) and also propose a new description method, edge uniformity. We evaluate the performance of the methods in the viewpoint of attentive object recognition.

Saliency-Guided Local Full-Reference Image Quality Assessment

“…Equation ( 8) implies that regions with strong edges are more salient to the human visual system than those with weak edges since edge information conveys essential information about the visual scene to humans [47]. A larger value of V(i) = max(E( f , i), E(g, i)) indicated a higher visual saliency at pixel location i.…”