A Visual Comfort Assessment Approach of Stereoscopic Images based on Random Forest Regressor

Visual comfort is an important component for stereoscopic image quality and viewing experience. In recent years, it has been widely investigated through feature analysis and different kinds of subjective experiments. In this paper, according to the existing works for objective evaluation model, we have screened and extracted thirteen disparity-based and content-based features through the stereoscopic image. To find the relationship between these features and the given mean opinion score(MOS) of each image, the method of Random Forest (RF) Regressor was used for simulation on the NBU S3D-VCA and IVY database. Compared with other state-of-the-art methods, the experimental results of two benchmarks data have confirmed the superior performance of our proposed approach.