Premium
8.4: Perceptual Quality Assessment for Motion Compensated Frame Interpolation
Author(s) -
Choi Kang Sun,
Yun YeoMin,
Han JongWoo,
Ko SungJea
Publication year - 2010
Publication title -
sid symposium digest of technical papers
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.351
H-Index - 44
eISSN - 2168-0159
pISSN - 0097-966X
DOI - 10.1889/1.3499824
Subject(s) - motion interpolation , interpolation (computer graphics) , metric (unit) , artificial intelligence , frame (networking) , computer vision , similarity (geometry) , computer science , inter frame , reference frame , motion (physics) , perception , correlation , mathematics , pattern recognition (psychology) , image (mathematics) , block matching algorithm , video tracking , object (grammar) , geometry , engineering , telecommunications , operations management , neuroscience , biology
In this paper, a novel nonreference video quality metric is proposed to evaluate motion compensated frame interpolation MCFI methods. Instead of using any exact reference frames, the proposed method estimates appropriate reference regions for every local region of interpolated frames by exploiting temporal correlation in frame interpolated video sequences. Then, the quality of the interpolated frame can be measured accurately by quantifying the image structural similarity between the corresponding local regions. Experimental results demonstrate that the proposed metric shows high correlation with the subjective evaluation of human observers significantly.