Abstract:Three-dimensional (3-D) visual comfort assessment (VCA) is a particularly important and challenging topic, which involves automatically predicting the degree of visual comfort in line with human subjective judgment. State-of-the-art VCA models typically focus on minimizing the distance between predicted visual comfort scores and subjective mean opinion scores (MOSs) by training a regression model. However, obtaining precise MOSs is often expensive and time-consuming, which greatly constrains the extension of existing MOS-aware VCA models. This study is inspired by the fact that humans tend to conduct a preference judgment between two stereoscopic images in terms of visual comfort. We propose to train a robust VCA model on a set of preference labels instead of MOSs. The preference label, representing the relative visual comfort of preference stereoscopic image pairs (PSIPs), is generally precise and can be obtained at much lower cost compared with MOS. More specifically, some representative stereoscopic images are first selected to generate the PSIP training set. Then, we use a support vector machine to learn a preference classification model by taking a differential feature vector and the corresponding preference label of each PSIP as input. Finally, given a testing sample, by considering a full-round paired comparison with all the selected representative stereoscopic images, the visual comfort score can be estimated via a simple linear mapping strategy. Experimental results on our newly built 3-D image database demonstrate that the proposed method can achieve a better performance compared with the models trained on MOSs. (C) The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

On Predicting Visual Comfort of Stereoscopic Images: A Learning to Rank Based Approach

Visual Comfort Assessment for Stereoscopic Images Based on Sparse Coding with Multi-Scale Dictionaries.

Leveraging Visual Attention and Neural Activity for Stereoscopic 3D Visual Comfort Assessment

Three-dimensional visual comfort assessment via preference learning

A Risk-Aware Pairwise Rank Learning Approach for Visual Discomfort Prediction of Stereoscopic 3D

Using Saliency-Weighted Disparity Statistics for Objective Visual Comfort Assessment of Stereoscopic Images

3D Visual Comfort Assessment Via Sparse Coding.

Hierarchical Visual Comfort Assessment for Stereoscopic Image Retargeting

A Visual Comfort Assessment Approach Of Stereoscopic Images Based On Random Forest Regressor

Assessment algorithm on visual comfort of stereoscopic image based on visual attention

A 3D Visual Comfort Metric Based on Binocular Asymmetry Factor.

An Ob jective Stereoscopic Image Visual Comfort Assessment Metric Based on Visual Important Regions

Study on Comfort Prediction of Stereoscopic Images Based on Improved Saliency Detection.

Prediction of visual discomfort of stereoscopic images based on saliency analysis

A Visual Comfort Assessment Metric for Stereoscopic Images.

A New 3D Video Database for Stereoscopic Video Comfortable Prediction

Hybrid Distortion Aggregated Visual Comfort Assessment for Stereoscopic Image Retargeting

Objective Visual Comfort Assessment Model of Stereoscopic Images Based on BP Neural Network.

Learning a No-Reference Quality Predictor of Stereoscopic Images by Visual Binocular Properties

Learning Receptive Fields and Quality Lookups for Blind Quality Assessment of Stereoscopic Images

Stereoscopic Image Discomfort Prediction Using Dual-Stream Multi-Level Interactive Network