Content-based subjective quality prediction in stereoscopic videos with machine learning

Abstract

A model exploiting machine learning and content analysis is proposed to predict the subjective quality of stereoscopic videos. This model offers an automated, accurate and consistent subjective quality prediction. The feasibility and accuracy of the proposed technique has been thoroughly analysed with extensive subjective experiments and simulations. Results illustrate that a performance measure of 0.954 in subjective quality prediction can be achieved with the proposed technique. © 2012 The Institution of Engineering and Technology.