Objective visual comfort assessment model of stereoscopic images based on BP neural network

The level of visual comfort human perceive for stereoscopic three-dimensional (S3D) images has been taking an important role in recent decades. Generally, the subjective assessment has been used to evaluate the visual comfort. However, it is time-consuming and man-powered experimental process. So how to develop effective and faithful objective visual comfort assessment (VCA) model is still a challenge. In the paper, we have proposed a new approach for objective evaluation based on the back propagation(BP) neural network. To establish the assessment model, the content and disparities related features of visual important regions (VIR) were both considered as the input. Experimental results showed high consistency with subject judgment compared with the state-of-the-art models on the same benchmark database.