Improved Retinex-Theory-Based Low-Light Image Enhancement Algorithm
Source abstract
Abstract
Researchers working on image processing have had a hard time handling low-light images due to their low contrast, noise, and brightness. This paper presents an improved method that uses the Retinex theory to enhance low-light images, with a network model mainly composed of a Decom-Net and an Enhance-Net. Residual connectivity is fully utilized in both the Decom-Net and Enhance-Net to reduce the possible loss of image details. Additionally, Enhance-Net introduces a positional pixel attention mechanism that directly incorporates the global information of the image. Specifically, Decom-Net serves to decompose the low-light image into illumination and reflection maps, and Enhance-Net serves to increase the brightness of the illumination map. Finally, via adaptive image fusion, the reflectance map and the enhanced illuminance map are fused to obtain the final enhanced image. Experiments show better results in terms of both subjective visual aspects and objective evaluation indicators. Compared to RetinexNet, the proposed method shows improvements in the full-reference evaluation metrics, including a 4.6% improvement in PSNR, a 1.8% improvement in SSIM, and a 10.8% improvement in LPIPS. Additionally, it achieved an average improvement of 17.3% in the no-reference evaluation metric NIQE.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Image Enhancement Techniques · Advanced Image Fusion Techniques · Advanced Image Processing Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Dalian Naval Academy, Dalian 116013, China
College of Information Science and Engineering, Northeastern University, Shenyang 110819, China
References (27)
Show source-provided references
1. A review on low light video image enhancement algorithms
2. Low-light image and video enhancement using deep learning: A survey10.1109/TPAMI.2021.3126387 ↗
3. Benchmarking low-light image enhancement and beyond10.1007/s11263-020-01418-8 ↗
4. Zheng, S., Ma, Y., Pan, J., Lu, C., and Gupta, G. (2022). Low-Light Image and Video Enhancement: A Comprehensive Survey and Beyond. arXiv.
5. Wei, C., Wang, W., Yang, W., and Liu, J. (2018). Deep retinex decomposition for low-light enhancement. arXiv.
6. R2rnet: Low-light image enhancement via real-low to real-normal network10.1016/j.jvcir.2022.103712 ↗
7. Adaptive histogram equalization and its variations10.1016/S0734-189X(87)80186-X ↗
8. Realization of the contrast limited adaptive histogram equalization (CLAHE) for real-time image enhancement10.1023/B:VLSI.0000028532.53893.82 ↗
9. Rahman, Z.-u., Jobson, D.J., and Woodell, G.A. (1996, January 19). Multi-scale retinex for color image enhancement. Proceedings of the 3rd IEEE International Conference on Image Processing, Lausanne, Switzerland.10.1109/ICIP.1996.560995 ↗
10. A multiscale retinex for bridging the gap between color images and the human observation of scenes10.1109/83.597272 ↗
11. LLNet: A deep autoencoder approach to natural low-light image enhancement10.1016/j.patcog.2016.06.008 ↗
12. LE-GAN: Unsupervised low-light image enhancement network using attention module and identity invariant loss10.1016/j.knosys.2021.108010 ↗
13. Ma, L., Ma, T., Liu, R., Fan, X., and Luo, Z. (2022, January 18–24). Toward fast, flexible, and robust low-light image enhancement. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA.10.1109/CVPR52688.2022.00555 ↗
14. Zhu, A., Zhang, L., Shen, Y., Ma, Y., Zhao, S., and Zhou, Y. (2020, January 6–10). Zero-shot restoration of underexposed images via robust retinex decomposition. Proceedings of the 2020 IEEE International Conference on Multimedia and Expo (ICME), London, UK.10.1109/ICME46284.2020.9102962 ↗
15. Lv, F., Lu, F., Wu, J., and Lim, C. (2018, January 3–6). MBLLEN: Low-Light Image/Video Enhancement Using CNNs. Proceedings of the BMVC, Newcastle, UK.
16. LightenNet: A convolutional neural network for weakly illuminated image enhancement10.1016/j.patrec.2018.01.010 ↗
17. Enlightengan: Deep light enhancement without paired supervision10.1109/TIP.2021.3051462 ↗
18. DSLR: Deep stacked Laplacian restorer for low-light image enhancement10.1109/TMM.2020.3039361 ↗
19. Zhang, L., Zhang, L., Liu, X., Shen, Y., Zhang, S., and Zhao, S. (2019, January 21–25). Zero-shot restoration of back-lit images using deep internal learning. Proceedings of the 27th ACM International Conference on Multimedia, Nice, France.10.1145/3343031.3351069 ↗
20. Learning to enhance low-light image via zero-reference deep curve estimation
21. Hore, A., and Ziou, D. (2010, January 23–26). Image quality metrics: PSNR vs. SSIM. Proceedings of the 2010 20th International Conference on Pattern Recognition, Istanbul, Turkey.10.1109/ICPR.2010.579 ↗
22. Study of subjective and objective quality assessment of video10.1109/TIP.2010.2042111 ↗
23. Naturalness preserved enhancement algorithm for non-uniform illumination images10.1109/TIP.2013.2261309 ↗
24. Li, C., Guo, C., Han, L., Jiang, J., Cheng, M.-M., Gu, J., and Loy, C.C. (2021). Lighting the darkness in the deep learning era. arXiv.
25. LIME: Low-light image enhancement via illumination map estimation10.1109/TIP.2016.2639450 ↗
26. Contrast enhancement based on layered difference representation of 2D histograms10.1109/TIP.2013.2284059 ↗
27. Guo, C., Li, C., Guo, J., Loy, C.C., Hou, J., Kwong, S., and Cong, R. (2020, January 14–19). Zero-reference deep curve estimation for low-light image enhancement. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.10.1109/CVPR42600.2020.00185 ↗