TY - GEN
T1 - Color barycenter model based multi-histogram mapping and merging for image enhancement
AU - Zhang, Qieshi
AU - Kamata, Sei Ichiro
N1 - Publisher Copyright:
© 2013; MVA Organization. All rights reserved.
PY - 2013
Y1 - 2013
N2 - In this paper, the color barycenter model (CBM) based image enhancement method using multihistogram mapping and merging is presented. Generally, histogram analysis based methods are effective for contrast enhancement, but this kind of method is hard to enhance the dark and bright regions efficiently simultaneously, such as the back-light image. To solve this problem, a mapping function is studied for multihistogram mapping to obtain several images with different contrast, and merging them by the best patch selecting of every position. Firstly, using the CBM to calculate the gray component as the input data. Secondly, obtaining several image with different contrast by our mapping function. Thirdly, calculating the gradient feature of the separated patches and selecting the best ones for merging. Finally, using the mix Gaussian filter to smooth the merged image. Based on the proposed approach, enhancement can be achieved for global/local regions under different light conditions. The experimental results show better effectiveness than other methods.
AB - In this paper, the color barycenter model (CBM) based image enhancement method using multihistogram mapping and merging is presented. Generally, histogram analysis based methods are effective for contrast enhancement, but this kind of method is hard to enhance the dark and bright regions efficiently simultaneously, such as the back-light image. To solve this problem, a mapping function is studied for multihistogram mapping to obtain several images with different contrast, and merging them by the best patch selecting of every position. Firstly, using the CBM to calculate the gray component as the input data. Secondly, obtaining several image with different contrast by our mapping function. Thirdly, calculating the gradient feature of the separated patches and selecting the best ones for merging. Finally, using the mix Gaussian filter to smooth the merged image. Based on the proposed approach, enhancement can be achieved for global/local regions under different light conditions. The experimental results show better effectiveness than other methods.
UR - http://www.scopus.com/inward/record.url?scp=85083079673&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85083079673&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85083079673
SN - 9784901122139
T3 - Proceedings of the 13th IAPR International Conference on Machine Vision Applications, MVA 2013
SP - 238
EP - 241
BT - Proceedings of the 13th IAPR International Conference on Machine Vision Applications, MVA 2013
PB - MVA Organization
T2 - 13th IAPR International Conference on Machine Vision Applications, MVA 2013
Y2 - 20 May 2013 through 23 May 2013
ER -