LPDM-YOLOv8: A Lightweight Low-Light Object Detection Method Based on Laplacian Pyramid

Yahao Guo; Dongxiang Fu

doi:10.71222/0tcmv848

Authors

Yahao Guo University of Shanghai for Science and Technology, Yangpu District, Shanghai 200093, China Author
Dongxiang Fu University of Shanghai for Science and Technology, Yangpu District, Shanghai 200093, China Author

DOI:

https://doi.org/10.71222/0tcmv848

Keywords:

YOLOv8, Laplacian pyramid, LPDM, brightness adaptation, multi-scale feature fusion, real-time object detection

Abstract

To address the performance degradation of YOLOv8 under low-light conditions, this paper proposes a lightweight low-light object detection method based on a Laplacian pyramid. A Laplacian Pyramid Dimming Module (LPDM) is integrated into YOLOv8 to construct an end-to-end detection framework with illumination-adaptive enhancement and multi-scale feature fusion. The proposed module performs brightness-aware adjustment and hierarchical detail reconstruction, effectively improving feature representation in dark scenes with negligible computational overhead. Experiments on the ExDark dataset show that LPDM-YOLOv8n achieves an mAP of 0.524, corresponding to an 8.0% relative improvement over the baseline YOLOv8n, while maintaining real-time performance at 32 FPS. Notably, only 16 additional parameters are introduced without increasing FLOPs. The results demonstrate that the proposed method significantly enhances detection robustness under low-light conditions while preserving efficiency, making it suitable for real-time applications.

References

1. C. Chen, Q. Chen, J. Xu, and V. Koltun, "Learning to see in the dark," In Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 3291-3300. doi: 10.1109/cvpr.2018.00347

2. J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You only look once: Unified, real-time object detection," In Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779-788. doi: 10.1109/cvpr.2016.91

3. Y. P. Loh, and C. S. Chan, "Getting to know low-light images with the exclusively dark dataset," Computer vision and image understanding, vol. 178, pp. 30-42, 2019. doi: 10.1016/j.cviu.2018.10.010

4. D. J. Jobson, Z. U. Rahman, and G. A. Woodell, "A multiscale retinex for bridging the gap between color images and the human observation of scenes," IEEE Transactions on Image processing, vol. 6, no. 7, pp. 965-976, 1997.

5. C. Guo, C. Li, J. Guo, C. C. Loy, J. Hou, S. Kwong, and R. Cong, "Zero-reference deep curve estimation for low-light image enhancement," In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 1780-1789.

6. X. Yin, Z. Yu, Z. Fei, W. Lv, and X. Gao, "Pe-yolo: Pyramid enhancement network for dark object detection," In International conference on artificial neural networks, September, 2023, pp. 163-174. doi: 10.1007/978-3-031-44195-0_14

7. T. Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, "Feature pyramid networks for object detection," In Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117-2125.