TY - JOUR
T1 - An optimization-based approach to gamma correction parameter estimation for low-light image enhancement
AU - Jeong, Inho
AU - Lee, Chul
N1 - Publisher Copyright:
© 2021, The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature.
PY - 2021/5
Y1 - 2021/5
N2 - We propose an efficient low-light image enhancement algorithm based on an optimization-based approach for gamma correction parameter estimation. We first separate an input color image into the luminance and chrominance channels, and then normalize the luminance channel using the logarithmic function to make it consistent with the human perception. Then, we divide the luminance image into dark and bright regions, and estimate the optimal gamma correction parameter for each region independently. Specifically, based on the statistical properties of the input image, we formulate a convex optimization problem that maximizes the image contrast subject to the constraint on the gamma value. By efficiently solving the optimization problems using the convex optimization theories, we obtain the optimal gamma parameter for each region. Finally, we obtain an enhanced image by merging the independently enhanced dark and bright regions with the optimal gamma parameters. Experimental results on real-world images demonstrate that the proposed algorithm can provide higher enhancement performance than state-of-the-art algorithms in terms of both subjective and objective evaluations, while providing a substantial improvement in speed.
AB - We propose an efficient low-light image enhancement algorithm based on an optimization-based approach for gamma correction parameter estimation. We first separate an input color image into the luminance and chrominance channels, and then normalize the luminance channel using the logarithmic function to make it consistent with the human perception. Then, we divide the luminance image into dark and bright regions, and estimate the optimal gamma correction parameter for each region independently. Specifically, based on the statistical properties of the input image, we formulate a convex optimization problem that maximizes the image contrast subject to the constraint on the gamma value. By efficiently solving the optimization problems using the convex optimization theories, we obtain the optimal gamma parameter for each region. Finally, we obtain an enhanced image by merging the independently enhanced dark and bright regions with the optimal gamma parameters. Experimental results on real-world images demonstrate that the proposed algorithm can provide higher enhancement performance than state-of-the-art algorithms in terms of both subjective and objective evaluations, while providing a substantial improvement in speed.
KW - Contrast enhancement
KW - Convex optimization
KW - Gamma correction
KW - Image fusion
KW - Low-light image enhancement
KW - Parameter estimation
UR - http://www.scopus.com/inward/record.url?scp=85101451997&partnerID=8YFLogxK
U2 - 10.1007/s11042-021-10614-8
DO - 10.1007/s11042-021-10614-8
M3 - Article
AN - SCOPUS:85101451997
SN - 1380-7501
VL - 80
SP - 18027
EP - 18042
JO - Multimedia Tools and Applications
JF - Multimedia Tools and Applications
IS - 12
ER -