Title :
An Evaluation Index Based on Parameter Weight for Image Inpainting Quality
Author :
Wang, Song ; Li, Hong ; Zhu, Xia ; Li, Ping
Author_Institution :
Sch. of Inf. Sci. & Eng., Central South Univ., Changsha
Abstract :
Traditional image quality evaluation index cannot explicitly reflect image inpainting quality. In order to make more reasonable measures on image inpainting, this paper proposed an evaluation method based on parameter weight through analyzing the reason for subjective distortion feeling after image inpainting. The new method synthetically evaluates image inpainting quality from three aspects: luminance, definition and gradient similarity. Experimental results demonstrate that the novel index can evaluate image inpainting quality more comprehensive and more similar to human eyes feeling than traditional index.
Keywords :
image processing; evaluation index; image inpainting quality; luminance; parameter weight; subjective distortion; Digital images; Distortion measurement; Equations; Eyes; Humans; Image analysis; Image quality; Image restoration; Information science; Pixel; Image quality evaluation index; image inpainting; parameter weight;
Conference_Titel :
Young Computer Scientists, 2008. ICYCS 2008. The 9th International Conference for
Conference_Location :
Hunan
Print_ISBN :
978-0-7695-3398-8
Electronic_ISBN :
978-0-7695-3398-8
DOI :
10.1109/ICYCS.2008.461