Title :
The image restoration method based on image segmentation and multiple feature fusion
Author_Institution :
Dept. of Comput. & Inf., Univ. of Fuyang Teachers, Fuyang, China
Abstract :
According to the lack of time consuming of global searching and considering only color error of image restoration algorithm based on the sample proposed by Criminisi, this paper proposes an improved image restoration method, which narrows search range of the optimal matching block according to segmentation pretreatment of the original image;and uses new matching criterion based on the color - structure - distance feature, through the fusion of D-S evidence theory,to determine the optimal matching block. The experimental result has showed the method not only can improve the image quality, but also can shorten the time of image restoration.
Keywords :
image colour analysis; image fusion; image matching; image restoration; image retrieval; image segmentation; inference mechanisms; D-S evidence theory; color error; global searching; image quality improvement; image restoration method; image segmentation pretreatment; matching criterion; multiple feature fusion; optimal matching block; Image segmentation; D-S evidence theory; Mean Shift; image restoration; image segmentation;
Conference_Titel :
Software Engineering and Service Science (ICSESS), 2013 4th IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4673-4997-0
DOI :
10.1109/ICSESS.2013.6615472