DocumentCode
3164011
Title
Remove meat image´s reflective areas with a level set constraint inpainting algorithm based on exemplar patch
Author
Jia, Yuan ; Liu, Pengcheng ; Niu, Sijie
Author_Institution
Sch. of Comput. Sci. & Technol., Southwest Univ. of Sci. & Technol., Mianyang, China
fYear
2011
fDate
8-10 Aug. 2011
Firstpage
4150
Lastpage
4154
Abstract
Because of moisture existing in meat, there would be reflective areas in the meat image under visible light, which lead to a loss of some muscle information. In order to improve the accuracy for obtaining meat image´s information, it is necessary to restore the reflective areas as much as possible. According to the feature of meat image, combine with the principle of image inpainting based on fast marching method, an improved image completion algorithm by exemplar patch is presented in this paper. First, binding the restoration order with level set distance and replacing Robert operator with Prewitt. Then, by using three factors such as direction, geometric distance and level set distance, known pixels in exemplar patch are weighted during the procedure of matching. In addition, the amount of scanning data is reduced in the process of searching the best matched patch through neighborhood matching. After that, to eliminate staircase effect, add α-cut Mean Filter in the process of inpainting. The experimental results show that the inpainting effect of improved algorithm is better than other algorithms present in this paper, and the method can not only repair the reflective areas of meat image, but also be useful for other images´ inpainting.
Keywords
food processing industry; food products; image matching; α-cut mean filter; Robert operator; exemplar patch; fast marching method; geometric distance; image completion algorithm; image matching; level set constraint inpainting algorithm; level set distance; meat image; muscle information; reflective areas; staircase effect; Filtering algorithms; Image restoration; Level set; Maintenance engineering; Mathematical model; PSNR; Visualization; fast marching method; image inpainting; level set; meat image; seam effect;
fLanguage
English
Publisher
ieee
Conference_Titel
Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC), 2011 2nd International Conference on
Conference_Location
Deng Leng
Print_ISBN
978-1-4577-0535-9
Type
conf
DOI
10.1109/AIMSEC.2011.6010087
Filename
6010087
Link To Document