New Stereoscopic Image Quality Assessment Metric Based on Three Dimensional-discrete Cosine Transform for 3d Media

This study presents a new stereoscopic image quality assessment metric based on three dimensional-discrete cosine transform (3D-DCT) for 3D media. In the proposed metric, the factors affecting stereoscopic perception are first combined to construct a three dimensional signal (called as 3D stack). 3D stack's coefficients are calculated in the 3D-DeI domain. Then, comparisons between reference and distorted stereoscopic images are made in terms of three important coefficients after perceptual weighting. Finally, the authors have get the objective assessment scores of the distorted stereoscopic images. Experimental results show that, the proposed metric outperforms relevant existing metrics in respect to correlation with the subjective evaluation and can predict the stereoscopic image quality well.