Semi-reference image quality metric for virtual view images in free-viewpoint television

摘要

A semi-reference image quality assessment metric based on similarity measurement for synthesizedvirtual viewpoint image (WI) in free-viewpoint television system (FTV) is proposed in this paper. Thekey point of the proposed metric is taking resemblant information between WI and its neighbor view imagesfor quality assessment to make our metric to be extended to multi-semi-reference image quality assessmenteasily. The proposed metric first extracts impact factors from image features, then combines animage synthesis technique and similarity functions, in which, disparity information are taken into accountfor registering the resemblant regions. Experiments are divided into three phases. Phase Ⅰ is to verify thevalidation of the proposed metric by taking impaired images and original reference into account. The experimentalresults show the agreement between evaluation scores and bio-characteristic of human visualsystem. Phase Ⅱ shows the accordance with Phase Ⅰ by taking neighbor view as reference. The proposedmetric can be taken as a full reference one to evaluate the image quality even though the original referenceis absent. Phase Ⅲ is then performed to evaluate the quality of ⅤⅥ. Evaluation scores in the experimentalresults are able to evaluate the quality of ⅤⅥ.