Stereoscopic Image Quality Assessment Based on Cyclopean Image
Kaixuan Lu, Wei Zhu · 2016
Image quality assessment is an important part inimage processing. With the stereoscopic video technologydevelopment and popularization, quality assessment ofstereoscopic images play an important role in three-dimensional(3D) multimedia applications. But the stereoscopic images qualityassessment (SIQA) is facing many new challenges, such as depthperception, virtual view synthesis and asymmetric stereocompression and so on. In this paper, a new Full-Reference (FR)stereoscopic image quality assessment method is proposed byconsidering the binocular image combining. Firstly, combinestereo pairs to a cyclopean image accounting for binocular rivalryand contrast sensitivity. Secondly, gradient image and luminanceimage are got, and the similarity indexes of them are computed.Finally, the similarity indexes are mixed together to get a singlequality score. The experimental results have demonstrated thatthe proposed metric is highly consistent with the subjective testscores.