LUMEN: Low-light Unified Multi-stage Enhancement Network to Improve RetinaFace-Based Face Detection
DOI:
https://doi.org/10.62411/faith.3048-3719-123Keywords:
Computer vision, Face detection, Low-light image enhancement, LUMEN, RetinaFace, Retinex, Surveillance systemsAbstract
Face detection in low-light conditions remains challenging due to underexposure, noise, and unstable contrast, which significantly degrade the performance of convolutional-based models. Conventional enhancement techniques, such as Histogram Equalization, Contrast-Limited Adaptive Histogram Equalization (CLAHE), and Low-Light Image Enhancement (LIME), often improve brightness but introduce visual artifacts and lack robustness for face detection. This study proposes Low-light Unified Multi-stage Enhancement Network (LUMEN), a lightweight multi-stage image enhancement pipeline designed as a preprocessing module to improve RetinaFace-based face detection under low-light conditions. LUMEN integrates Multiscale Retinex with Color Restoration, adaptive gamma correction, CLAHE-based local contrast enhancement, controlled image fusion, and Non-Local Means denoising to jointly stabilize illumination, preserve texture, and maintain visual naturalness. The method is evaluated on a low-light subset of the Human Faces Object Detection Dataset using RetinaFace. Detection performance is assessed using detection rate and confidence score, while visual quality is evaluated using no-reference metrics such as the Natural Image Quality Evaluator (NIQE) and the Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE). Experimental results show that LUMEN achieves a face detection rate of 91% with a high confidence score (0.9545), rep-resenting an improvement of 33 percentage points over raw low-light images (58%), while maintaining detection performance comparable to CLAHE, which ranks as the second-best method, LUMEN delivers superior perceptual quality, evidenced by the lowest BRISQUE score, as well as more stable visual appearance with reduced noise amplification and fewer contrast artifacts in low-light facial regions. LUMEN also achieves the lowest BRISQUE score, indicating superior texture preservation and perceptual stability. Ablation studies confirm that Retinex and CLAHE are the most critical components for detection robustness, while gamma correction, fusion, and denoising mainly contribute to visual naturalness. These results demonstrate that LUMEN provides an effective and practical preprocessing solution for low-light face detection.
Downloads
References
W. Zhao, R. Chellappa, P. J. Phillips, and A. Rosenfeld, “Face recognition,” ACM Comput. Surv., vol. 35, no. 4, pp. 399–458, Dec. 2003, doi: 10.1145/954339.954342.
W. Liu, P. Zhao, X. Song, and B. Zhang, “A Survey of Low-light Image Enhancement,” Front. Comput. Intell. Syst., vol. 1, no. 3, pp. 88–92, Oct. 2022, doi: 10.54097/fcis.v1i3.2242.
L. R. Zuama, D. R. I. M. Setiadi, A. Susanto, S. Santosa, H.-S. Gan, and A. A. Ojugo, “High-Performance Face Spoofing Detection using Feature Fusion of FaceNet and Tuned DenseNet201,” J. Futur. Artif. Intell. Technol., vol. 1, no. 4, pp. 385–400, Feb. 2025, doi: 10.62411/faith.3048-3719-62.
R. M. Azeez, I. A. Alshabeeb, and W. M. R. Shakir, “Enhanced Face Recognition Using Dolphin Swarm Optimization with Eu-clidean Classification and PCA,” J. Futur. Artif. Intell. Technol., vol. 2, no. 3, pp. 405–416, Sep. 2025, doi: 10.62411/faith.3048-3719-127.
M. Alansari, O. A. Hay, S. Javed, A. Shoufan, Y. Zweiri, and N. Werghi, “GhostFaceNets: Lightweight Face Recognition Model From Cheap Operations,” IEEE Access, vol. 11, pp. 35429–35446, 2023, doi: 10.1109/ACCESS.2023.3266068.
D. J. Jobson, Z. Rahman, and G. A. Woodell, “A multiscale retinex for bridging the gap between color images and the human ob-servation of scenes,” IEEE Trans. Image Process., vol. 6, no. 7, pp. 965–976, Jul. 1997, doi: 10.1109/83.597272.
K. Zhang, W. Zuo, Y. Chen, D. Meng, and L. Zhang, “Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising,” IEEE Trans. Image Process., vol. 26, no. 7, pp. 3142–3155, Jul. 2017, doi: 10.1109/TIP.2017.2662206.
W. Yang, S. Wang, Y. Fang, Y. Wang, and J. Liu, “From Fidelity to Perceptual Quality: A Semi-Supervised Approach for Low-Light Image Enhancement,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, pp. 3060–3069. doi: 10.1109/CVPR42600.2020.00313.
X. Guo, “LIME: A Method for Low-light IMage Enhancement,” in Proceedings of the 24th ACM international conference on Multimedia, Oct. 2016, pp. 87–91. doi: 10.1145/2964284.2967188.
D. Tete and C. S. Panda, “A Review Paper on LIME: Low-Light Image Enhancement,” Int. J. Nov. Res. Dev., vol. 7, no. 12, 2022.
S. M. Pizer et al., “Adaptive histogram equalization and its variations,” Comput. Vision, Graph. Image Process., vol. 39, no. 3, pp. 355–368, Sep. 1987, doi: 10.1016/S0734-189X(87)80186-X.
H. Han, S. Shan, L. Qing, X. Chen, and W. Gao, “Lighting Aware Preprocessing for Face Recognition across Varying Illumination,” in Lecture Notes in Computer Science, 2010, pp. 308–321. doi: 10.1007/978-3-642-15552-9_23.
S. Rahman, M. M. Rahman, M. Abdullah-Al-Wadud, G. D. Al-Quaderi, and M. Shoyaib, “An adaptive gamma correction for image enhancement,” EURASIP J. Image Video Process., vol. 2016, no. 1, p. 35, Dec. 2016, doi: 10.1186/s13640-016-0138-1.
Z. Zhang, H. Zheng, R. Hong, M. Xu, S. Yan, and M. Wang, “Deep Color Consistent Network for Low-Light Image Enhancement,” in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2022, pp. 1889–1898. doi: 10.1109/CVPR52688.2022.00194.
H. Ackar, A. A. Almisreb, and M. A. Saleh, “A Review on Image Enhancement Techniques,” Southeast Eur. J. Soft Comput., vol. 8, no. 1, Apr. 2019, doi: 10.21533/scjournal.v8i1.175.
J. Wang, H. Wang, Y. Sun, and J. Yang, “Improved Retinex-Theory-Based Low-Light Image Enhancement Algorithm,” Appl. Sci., vol. 13, no. 14, p. 8148, Jul. 2023, doi: 10.3390/app13148148.
J. Deng, J. Guo, Y. Zhou, J. Yu, I. Kotsia, and S. Zafeiriou, “RetinaFace: Single-stage Dense Face Localisation in the Wild,” arXiv. May 04, 2019. [Online]. Available: http://arxiv.org/abs/1905.00641
A. M. Reza, “Realization of the Contrast Limited Adaptive Histogram Equalization (CLAHE) for Real-Time Image Enhancement,” J. VLSI signal Process. Syst. signal, image video Technol., vol. 38, no. 1, pp. 35–44, Aug. 2004, doi: 10.1023/B:VLSI.0000028532.53893.82.
E. H. Land, “The Retinex Theory of Color Vision,” Sci. Am., vol. 237, no. 6, pp. 108–128, Dec. 1977, doi: 10.1038/scientificamerican1277-108.
X. Guo, Y. Li, and H. Ling, “LIME: Low-Light Image Enhancement via Illumination Map Estimation,” IEEE Trans. Image Process., vol. 26, no. 2, pp. 982–993, Feb. 2017, doi: 10.1109/TIP.2016.2639450.
X. Pan, C. Li, Z. Pan, J. Yan, S. Tang, and X. Yin, “Low-Light Image Enhancement Method Based on Retinex Theory by Improving Illumination Map,” Appl. Sci., vol. 12, no. 10, 2022, doi: 10.3390/app12105257.
Z. Gu, M. Ju, and D. Zhang, “A novel Retinex image enhancement approach via brightness channel prior and change of detail prior,” Pattern Recognit. Image Anal., vol. 27, no. 2, pp. 234–242, Apr. 2017, doi: 10.1134/S1054661817020055.
C. Guo et al., “Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, pp. 1777–1786. doi: 10.1109/CVPR42600.2020.00185.
I. Majid Mohammed and N. Ashidi Mat Isa, “Contrast Limited Adaptive Local Histogram Equalization Method for Poor Contrast Image Enhancement,” IEEE Access, vol. 13, pp. 62600–62632, 2025, doi: 10.1109/ACCESS.2025.3558506.
J. Deng, J. Guo, E. Ververas, I. Kotsia, and S. Zafeiriou, “RetinaFace: Single-Shot Multi-Level Face Localisation in the Wild,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, pp. 5202–5211. doi: 10.1109/CVPR42600.2020.00525.
B. B. Singh and S. Patel, “Efficient Medical Image Enhancement using CLAHE Enhancement and Wavelet Fusion,” Int. J. Comput. Appl., vol. 167, no. 5, pp. 1–5, Jun. 2017, doi: 10.5120/ijca2017913277.
J. Zhong, H. Xie, and F. L. Wang, “The research trends in recommender systems for e-learning,” Asian Assoc. Open Univ. J., vol. 14, no. 1, pp. 12–27, Nov. 2019, doi: 10.1108/AAOUJ-03-2019-0015.
C. Li, J. Zhu, L. Bi, W. Zhang, and Y. Liu, “A low-light image enhancement method with brightness balance and detail preservation,” PLoS One, vol. 17, no. 5, p. e0262478, May 2022, doi: 10.1371/journal.pone.0262478.
W. Kim, “Low-Light Image Enhancement: A Comparative Review and Prospects,” IEEE Access, vol. 10, pp. 84535–84557, 2022, doi: 10.1109/ACCESS.2022.3197629.
D. Mishra and B. Palkar, “Image Fusion Techniques: A Review,” Int. J. Comput. Appl., vol. 130, no. 9, pp. 7–13, Nov. 2015, doi: 10.5120/ijca2015907084.
K. Han, Y. Wang, Q. Tian, J. Guo, C. Xu, and C. Xu, “GhostNet: More Features From Cheap Operations,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, pp. 1577–1586. doi: 10.1109/CVPR42600.2020.00165.
S. Fu, Q. Zhao, H. Liu, Q. Tao, and D. Liu, “Low-light object detection via adaptive enhancement and dynamic feature fusion,” Alexandria Eng. J., vol. 126, pp. 60–69, Jul. 2025, doi: 10.1016/j.aej.2025.04.047.
M. F. Khan, D. Goyal, M. M. Nofal, E. Khan, R. Al-Hmouz, and E. Herrera-Viedma, “Fuzzy-Based Histogram Partitioning for Bi-Histogram Equalisation of Low Contrast Images,” IEEE Access, vol. 8, pp. 11595–11614, 2020, doi: 10.1109/ACCESS.2020.2965174.
H. Wang and Y. Song, “Adaptive Low Light Enhancement via Joint Global-Local Illumination Adjustment,” arXiv. Apr. 01, 2025. [Online]. Available: http://arxiv.org/abs/2504.00400
X. Liang, X. Chen, K. Ren, X. Miao, Z. Chen, and Y. Jin, “Low-light image enhancement via adaptive frequency decomposition network,” Sci. Rep., vol. 13, no. 1, p. 14107, Aug. 2023, doi: 10.1038/s41598-023-40899-8.
S. Fu, J. Zhang, and Y. Luo, “A Retinex-based variational model for low-light image enhancement with noise transformation,” Pattern Recognit., vol. 172, p. 112781, Apr. 2026, doi: 10.1016/j.patcog.2025.112781.
Z. Liu et al., “Illuminating the Shadows: Enhanced Low-Light Image via a Retinex-based Model with Color Equalization,” Expert Syst. Appl., vol. 296, p. 129039, Jan. 2026, doi: 10.1016/j.eswa.2025.129039.
R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th ed. Pearson, 2018.
A. Setyono and D. Setiadi, “Robust Video Watermarking using Tchebichef Transform and Singular Value Decomposition on the Selected Frame Based YCbCr Color Space,” Int. J. Intell. Eng. Syst., vol. 13, no. 6, pp. 432–441, Dec. 2020, doi: 10.22266/ijies2020.1231.38.
Z. Hong et al., “Efficient Global Color, Luminance, and Contrast Consistency Optimization for Multiple Remote Sensing Images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 16, pp. 622–637, 2023, doi: 10.1109/JSTARS.2022.3229392.
M. S. Ragavendirane and S. Dhanasekar, “Low-Light Image Enhancement via New Intuitionistic Fuzzy Generator-Based Retinex Approach,” IEEE Access, vol. 13, pp. 38454–38469, 2025, doi: 10.1109/ACCESS.2025.3545258.
J. J. Jeon, J. Y. Park, and I. K. Eom, “Low-light image enhancement using gamma correction prior in mixed color spaces,” Pattern Recognit., vol. 146, p. 110001, Feb. 2024, doi: 10.1016/j.patcog.2023.110001.
S. A. Mim, J. Cheng, and J. Zhang, “Fusion-based framework for low-light image enhancement using BM3D filter and adaptive gamma correction,” Opt. Laser Technol., vol. 192, p. 113662, Dec. 2025, doi: 10.1016/j.optlastec.2025.113662.
L. Pi, Y. Mu, and L. Pi, “Multiscale processing of image gradient domain-based convolutional neural networks for visual feature enhancement,” J. Comb. Math. Comb. Comput., vol. 127a, Apr. 2025, doi: 10.61091/jcmcc127a-269.
D. Mamieva, A. B. Abdusalomov, M. Mukhiddinov, and T. K. Whangbo, “Improved Face Detection Method via Learning Small Faces on Hard Images Based on a Deep Learning Approach,” Sensors, vol. 23, no. 1, p. 502, Jan. 2023, doi: 10.3390/s23010502.
S. Santoso, D. R. I. M. Setiadi, and R. A. Pramunendar, “Comparative No-Reference Evaluation of Classical Image Sharpening Techniques under Varying Degradation Conditions,” J. Appl. Informatics Comput., vol. 9, no. 6, pp. 3442–3453, 2025, doi: 10.30871/jaic.v9i6.11430.
N. Yuan et al., “Low-Light Image Enhancement by Combining Transformer and Convolutional Neural Network,” Mathematics, vol. 11, no. 7, p. 1657, Mar. 2023, doi: 10.3390/math11071657.
J. Vazquez-Corral, G. D. Finlayson, and L. Herranz, “Improving the perception of low-light enhanced images,” Opt. Express, vol. 32, no. 4, p. 5174, Feb. 2024, doi: 10.1364/OE.509713.
A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-Reference Image Quality Assessment in the Spatial Domain,” IEEE Trans. Image Process., vol. 21, no. 12, pp. 4695–4708, Dec. 2012, doi: 10.1109/TIP.2012.2214050.
X. Wu et al., “Low-Light Enhancement Effect on Classification and Detection: An Empirical Study,” arXiv. Sep. 22, 2024. [Online]. Available: http://arxiv.org/abs/2409.14461
J. Zhang, C. Hou, X. Yang, X. Yang, W. Yang, and H. Cui, “Advancing face detection efficiency: Utilizing classification networks for lowering false positive incidences,” Array, vol. 22, p. 100347, Jul. 2024, doi: 10.1016/j.array.2024.100347.
T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, “Focal Loss for Dense Object Detection,” ArXiv. Feb. 07, 2018. [Online]. Available: http://arxiv.org/abs/1708.02002
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Rivaldi Ramadani, De Rosal Ignatius Moses Setiadi, Imanuel Harkespan, Kristoko Dwi Hartomo, Christian Arthur
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
