TY - GEN
T1 - SELF-SUPERVISED LEARNING ON A LIGHTWEIGHT LOW-LIGHT IMAGE ENHANCEMENT MODEL WITH CURVE REFINEMENT
AU - Wu, Wanyu
AU - Wang, Wei
AU - Jiang, Kui
AU - Xu, Xin
AU - Hu, Ruimin
N1 - Publisher Copyright:
© 2022 IEEE
PY - 2022
Y1 - 2022
N2 - Deep learning networks with deeper layers become a trend for their good performance but lacks the potential for real-time mobile deployment. Another challenge for paired training networks is the limited generalization capacity caused by the sample bias. To overcome these two challenges, we propose a lightweight self-supervised low-light image enhancement method, that trains with low light images only. Specifically, our method consists of a low-resolution dense CNN network stream and a full-resolution guidance stream, responsible for image-to-curve transformation with refinement and spatial guidance fusion, respectively. Then, a new self-supervised loss function is introduced to measure the restored patch-based color deviations among color channels. Experimental results show that our method gives competitive performance to the full-supervised approaches.
AB - Deep learning networks with deeper layers become a trend for their good performance but lacks the potential for real-time mobile deployment. Another challenge for paired training networks is the limited generalization capacity caused by the sample bias. To overcome these two challenges, we propose a lightweight self-supervised low-light image enhancement method, that trains with low light images only. Specifically, our method consists of a low-resolution dense CNN network stream and a full-resolution guidance stream, responsible for image-to-curve transformation with refinement and spatial guidance fusion, respectively. Then, a new self-supervised loss function is introduced to measure the restored patch-based color deviations among color channels. Experimental results show that our method gives competitive performance to the full-supervised approaches.
KW - image-to-curve transformation
KW - lightweight self-supervised network
KW - real-time low-light image enhancement
UR - https://www.scopus.com/pages/publications/85131242911
U2 - 10.1109/ICASSP43922.2022.9746348
DO - 10.1109/ICASSP43922.2022.9746348
M3 - 会议稿件
AN - SCOPUS:85131242911
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 1890
EP - 1894
BT - 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2022
Y2 - 22 May 2022 through 27 May 2022
ER -