Person Detection in Low-Light Environments: Evaluation of an Annotation-Free Approach
Keywords:
Low Light Environments, Dataset Construction, Data Augmentation, Image Enhancement, Object Detection, Deep LearningAbstract
This work investigates how to modify an existing dataset to enhance the capability of a predictive model' to detect objects in low-light environments. From an initial set of images captured under daylight conditions, synthetic darkened versions were generated. Subsequently, various training scenarios were organized to evaluate the impact of image transformation techniques and data augmentation strategies on predictive performance. The experiments performed demonstrate that the combination of original images and artificially darkened versions, coupled with data augmentation and neutral background images, results in more stable and accurate models for low-light detection, with negligible performance degradation in visible-light scenarios. The results highlight that it is possible to improve detector performance in real-world low-light settings without the need for new image acquisitions or additional annotation.
Downloads
References
Krizhevsky, A.; Sutskever, I.; Hinton, G.E., Imagenet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems. pp. 1097–1105, 2012.
Romera, Eduardo; Bergasa, Luis M.; Yang, Kailun; Alvarez, Jose M.; Barea1, Rafael. Bridging the Day and Night Domain Gap for Semantic Segmentation. In: IEEE Intelligent Vehicles Symposium, 2019, doi: 10.1109/IVS.2019.8813888
Xueqing Deng; Peng Wang; Xiaochen Lian; Shawn Newsam. NightLab: A Dual-level Architecture with Hardness Detection for Segmentation at Night. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.16917-16927, 2022, doi: 10.48550/arXiv.2204.05538.
Loh, Yuen Peng; Chan, Chee Seng. Getting to know low-light images with the Exclusively Dark dataset. In: Computer Vision and Image Understanding, v. 178, pp.30–42, 2019, doi: 10.1016/j.cviu.2018.10.010
Zongjiang Gao; Yingjun Zhang; Yuankui Li. Extracting features from infrared images using convolutional neural networks and transfer learning, In: Infrared Physics & Technology, v. 105, pp.103237, 2020, doi: 10.1016/j.infrared.2020.103237.
A. Kapoor; R. Greiner, Learning and classifying under hard budgets, In: Proceedings of the 16th European Conference on Machine Learning, Springer-Verlag, Berlin, Heidelberg, pp. 170–181, 2005, doi: 10.1007/11564096_20.
P. Melville; M. Saar-Tsechansky; F. Provost; R. Mooney, An Expected utility approach to active feature-value acquisition, In: Proceedings of the Fifth IEEE International Conference on Data Mining (ICDM ׳05), IEEE Computer Society. Washington, DC, USA, pp. 745–748, 2005, doi: 10.1109/ICDM.2005.23.
Ma, Xiang; Xia, Lei; Zhou, Bolei; et al. The Effect of Improving Annotation Quality on Object Detection Datasets. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). 2022, doi: 10.1109/CVPRW56347.2022.00532.
C. Agnew, A. Scanlan, P. Denny, E. M. Grua, P. van de Ven and C. Eising. Annotation Quality Versus Quantity for Object Detection and Instance Segmentation, IEEE Access, v. 12, pp. 140958-140977, 2024, doi: 10.1109/ACCESS.2024.3467008
Loh, Yuen Peng and Chan, Chee Seng. Getting to Know Low-light Images with The Exclusively Dark Dataset. Computer Vision and Image Understanding, v. 178, pp. 30-42, 2019, doi: 10.1016/j.cviu.2018.10.010.
Ren, B., Xu, Z., Zhao, J. et al. Complex dark environment-oriented object detection method based on YOLO-AS. Sci Rep 15, 21873, 2025, doi: 10.1038/s41598-025-07348-0.
Li H, Fu Y. FCDFusion: A fast, low color deviation method for fusing visible and infrared image pairs. Computational Visual Media,11(1): pp.195-211, 2025, doi: 10.26599/CVM.2025.9450330.
Donia, Eman; El-Rabaie, El-Sayed; El-Samie, Fathi; Faragallah, Osama; A.El-Hag, Noha. Infrared image fusion for quality enhancement. Journal of Optics. 52. Pp.1-7, 2023, doi: 10.1007/s12596-022-01018-4.
Haoran Zhu; Wenying Zhang. Infrared image fusion for quality enhancement. IEEE Access, v. 13, pp. 61862-61875, 2025, doi: 10.1109/ACCESS.2025.3557799.
H. Kim, J. Kim and J. Kim. Image-to-Image Translation for Near-Infrared Image Colorization, 2022 International Conference on Electronics, Information, and Communication (ICEIC), Jeju, Korea, Republic of, 2022, pp. 1-4, 2023, doi: 10.1109/ICEIC54506.2022.9748773.
D. Ma, S. Li, J. Su, Y. Xian and T. Zhang, "Visible-to-Infrared Image Translation for Matching Tasks," in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 17, pp. 18199-18213, 2024, doi: 10.1109/JSTARS.2024.3468456.
Jiawei Guo; Jieming Ma; Ángel F. García-Fernández; Yungang Zhang; Haining Liang. A Survey on Image Enhancement for Low-Light Images. Heliyon, v. 9, n. 4, 2023, doi: 10.1016/j.heliyon.2023.e14558.
Y. Luo, X. Chen, J. Ling, C. Huang, W. Zhou and G. Yue. Unsupervised Low-Light Image Enhancement With Self-Paced Learning, in IEEE Transactions on Multimedia, vol. 27, pp. 1808-1820, 2025, doi: 10.1109/TMM.2024.3521752.
Saifullah S, Dreżewski R. Advanced Medical Image Segmentation Enhancement: A Particle-Swarm-Optimization-Based Histogram Equalization Approach, Applied Sciences, v.14, n. 2, pp. 923-949, 2024, doi: 10.3390/app14020923
R. Maini and H. Aggarwel, “A Comprehensive Review of Image Enhancement Techniques,” Journal of Computing, Vol. 2, No. 3, 2010, pp. 8-13. Available at: https://arxiv.org/pdf/1003.4053. Accessed: 15 jul. 2025.
Mohammad Mahdizadeh, Shijie Chen, Peng Ye. Illuminating the night: A light source-aware and exposure-balanced low-light enhancement approach for real nighttime scenes, Digital Signal Processing, v. 159, pp. 104999, 2025, doi: 10.1016/j.dsp.2025.104999.
Gonzalez, Rafael C.; Woods, Richard E. Digital Image Processing. 3. ed. Pearson, ISBN 978-0131687288, 2018.
Syl Arena. Speedliter’s Handbook: Learning to Craft Light with Canon Speedlites, ed.2, Peachpit Press, ISBN 978-0-134-00791-5, 2016.
ULTRALYTICS. YOLOv5 by Ultralytics. 2020. Available at : https://docs.ultralytics.com/models/yolov5/. Accessed: 13 abr. 2025.
Gomes, Isabelli Pinto. Github Repository: YOLOv5 by Ultralytics. 2020. Available at: https://github.com/IsabelliGomes/low-light-person-detection-yolo. Accessed: 01 jun. 2026.
Chongyi Li; Chunle Guo; Linghao Han; Jun Jiang; Ming-Ming Cheng; Jinwei Gu. Low-Light Image and Video Enhancement Using Deep Learning: A Survey. In: IEEE Transactions on Pattern Analysis and Machine Intelligence , v.44, n.12, pp. 9396-9416, 2022, doi: 10.1109/TPAMI.2021.3126387.
