PROJECT TITLE :

A Biologically Inspired Automatic System for Media Quality Assessment

ABSTRACT:

Photo aesthetic quality evaluation could be a challenging task in artificial intelligence systems. In this paper, we propose a biologically impressed aesthetic descriptor that mimicks humans sequentially perceiving visually/semantically salientIn general, visually salient regions are perceived by low-level visual features, such as the high distinction between the foreground and the background objects; while semantically salient regions are perceived by high-level visual options such as human faces.regions in a very photo. In specific, a weakly supervised learning paradigm is developed to project the local image descriptors into a low-dimensional semantic area. Then, every graphlet will be described by multiple sorts of visual options, each in low-level and in high-level. Since humans typically understand solely a few salient regions in an exceedingly photo, a sparsity-constrained graphlet ranking algorithm is proposed that seamlessly integrates both the low-level and also the high-level visual cues. Top-ranked graphlets are those visually/semantically distinguished native aesthetic descriptors during a photo. They are sequentially linked into a path that simulates humans actively viewing method. Finally, we have a tendency to learn a probabilistic aesthetic live based on such actively viewing paths (AVPs) from the training photos. Experimental results show that: 1) the AVPs are eighty seven.65% in keeping with real human gaze shifting paths, as verified by the eye-tracking information and a couple of) our aesthetic measure outperforms many of its competitors.