Objective quality assessment of stereo images based on ICA and BT-SVM
Jincui Cheng, Sumei Li · 2012
With more and more applications of stereo information technology, the quality assessment of stereo image is quite needed. However, it is very difficult to find an assessment metric which really denotes the quality feature of stereo image. In this paper, a novel assessment method is proposed based on independent component analysis (ICA) and binary tree support vector machine (BT-SVM). Firstly, a set of independent basis images are extracted by ICA; then, the BT-SVM is used as a quality grade classifier to judge the grade of the tested stereo image. Experimental results demonstrate that the quality assessment results based upon the proposed metric are consistent with those obtained by subjective assessment and its correct classification ratio is more than 90%.