A Two-Stage Framework for Object Detection in Low-Light Images Using Image Enhancement and Deep Learning Models
DOI:
https://doi.org/10.47134/jtsi.v3i1.5349Keywords:
Low-Light Images, Object Detection, Image Enhancement, Yolov9, Faster R-CNN, Deep LearningAbstract
In low-lighting scenarios in object detection, a major challenge exists owing to reduced lighting, greater noise, lower contrast, and lighting changes. Thus, such scenarios have a significant effect on vision-based systems used in surveillance, path detection for autonomous vehicles, and security surveillance. A two-tier method using classical image processing and a deep learning platform for object detection in images is proposed and implemented in this work. The first stage uses a dedicated image processing chain aimed at increasing image brightness, contrast, and clarity while eliminating image noise. These processed images are then subjected to evaluation by two separate object detection models: YOLOv9 and Faster R-CNN. From ExDark dataset testing, the effectiveness of the method implemented has a mean Average Precision value of 96% at IOU= 0.50 for YOLOv9 and 88% mAP@50 for Faster R-CNN.
References
Anil, S., & Le, T. (2018). CNN-based learning using reflection and Retinex models for intrinsic image decomposition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 6674–6683).
Banik, P. P., Saha, R., & Kim, K.-D. (2018). Contrast enhancement of low-light image using histogram equalization and illumination adjustment. In 2018 International Conference on Electronics, Information, and Communication (ICEIC) (pp. 1–4). IEEE. https://doi.org/10.23919/ELINFocom.2018.8330564
Bertalmío, M., Caselles, V., & Provenzi, E. (2009). Issues about Retinex theory and contrast enhancement. International Journal of Computer Vision, 83(1), 101–119.
Chaturvedi, A., Khanna, R., & Kumar, V. (2016). An analysis of region growing image segmentation schemes. International Journal of Computer Trends and Technology, 34(1), 46–51. https://doi.org/10.14445/22312803/IJCTT-V34P108
Chen, W., & Shah, T. (2021). Exploring low-light object detection techniques. arXiv.
Fraser, D., Schowengerdt, R. A., & Briggs, I. (1984). Rectification of multichannel images in mass storage using image transposition. Computer Vision, Graphics, and Image Processing, 27(1), 126. https://doi.org/10.1016/0734-189X(84)90091-4
Fu, G., et al. (2023). Learning adaptive feature for object detection in low-light conditions. SSRN. https://doi.org/10.2139/ssrn.4625592
Gao, X., Zhang, M., & Luo, J. (2022). Low-light image enhancement via Retinex-style decomposition of denoised deep image prior. Sensors, 22(15), 5593. https://doi.org/10.3390/s22155593
Gonzalez, R. C., Woods, R. E., & Masters, B. R. (2009). Digital image processing (3rd ed.). Journal of Biomedical Optics, 14(2), 029901. https://doi.org/10.1117/1.3115362
Guo, C., et al. (2020). Zero-reference deep curve estimation for low-light image enhancement. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE. https://doi.org/10.1109/CVPR42600.2020.00185
Guo, J., Ma, J., García-Fernández, Á. F., Zhang, Y., & Liang, H. (2023). A survey on image enhancement for low-light images. Heliyon, 9(4), e14558. https://doi.org/10.1016/j.heliyon.2023.e14558
Hai, J., et al. (2023). R2RNet: Low-light image enhancement via real-low to real-normal network. Journal of Visual Communication and Image Representation, 90, 103712. https://doi.org/10.1016/j.jvcir.2022.103712
Jiang, Y., et al. (2021). EnlightenGAN: Deep light enhancement without paired supervision. IEEE Transactions on Image Processing, 30, 2340–2349. https://doi.org/10.1109/TIP.2021.3051462
Le, H., & Samaras, D. (2019). Shadow removal via shadow image decomposition. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV) (pp. 8577–8586). IEEE. https://doi.org/10.1109/ICCV.2019.00867
Lee, C., Lee, C., & Kim, C.-S. (2012). Contrast enhancement based on layered difference representation. In 2012 IEEE International Conference on Image Processing. IEEE. https://doi.org/10.1109/ICIP.2012.6467022
Loh, Y. P., & Chan, C. S. (2019). Getting to know low-light images with the exclusively dark dataset. Computer Vision and Image Understanding, 178, 30–42. https://doi.org/10.1016/j.cviu.2018.10.010
Nie, F., Li, Z., Wang, R., & Li, X. (2023). An effective and efficient algorithm for k-means clustering with new formulation. IEEE Transactions on Knowledge and Data Engineering, 35(4), 3433–3443. https://doi.org/10.1109/TKDE.2022.3155450
Popev, A. T. (1995). Morphological operations on fuzzy sets. In Fifth International Conference on Image Processing and Its Applications (pp. 837–840). IEE. https://doi.org/10.1049/cp:19950778
Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149.
Ren, W., et al. (2018). Gated fusion network for single image dehazing. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE. https://doi.org/10.1109/CVPR.2018.00343
Tang, H., Zhu, H., Fei, L., Wang, T., Cao, Y., & Xie, C. (2023). Low-illumination image enhancement based on deep learning techniques: A brief review. Photonics, 10(2), 198. https://doi.org/10.3390/photonics10020198
Tarel, J.-P., & Hautière, N. (2009). Fast visibility restoration from a single color or gray-level image. In 2009 IEEE 12th International Conference on Computer Vision. IEEE. https://doi.org/10.1109/ICCV.2009.5459251
Thepade, S. D., Ople, M., Mahindra, V., Kulye, V., & Jamdar, S. (2022). Low light image contrast enhancement using blending of histogram equalization based methods DSIHE and MMBEBHE. In 2022 2nd Asian Conference on Innovation in Technology (ASIANCON). IEEE. https://doi.org/10.1109/ASIANCON55314.2022.9909318
Vijayalakshmi, D., Nath, M. K., & Acharya, O. P. (2020). A comprehensive survey on image contrast enhancement techniques in spatial domain. Signal, Image and Video Processing, 21(1). https://doi.org/10.1007/s11220-020-00305-3
Vinoth, K., & S. P. (2024). Lightweight object detection in low light: Pixel-wise depth refinement and TensorRT optimization. Results in Engineering, 23, 102510. https://doi.org/10.1016/j.rineng.2024.102510
Wu, S., Wang, S., Laganiere, R., Liu, C., Wong, H.-S., & Xu, Y. (2018). Exploiting target data to learn deep convolutional networks for scene-adapted human detection. IEEE Transactions on Image Processing, 27(3), 1418–1432. https://doi.org/10.1109/TIP.2017.2779271
Zhang, B., Zhu, G., & Zhu, Z. (2020). A TV-log nonconvex approach for image deblurring with impulsive noise. Signal Processing, 174, 107631. https://doi.org/10.1016/j.sigpro.2020.107631
Zuiderveld, K. (1994). Contrast limited adaptive histogram equalization. In A. S. Glassner (Ed.), Graphics gems (pp. 474–485). Elsevier. https://doi.org/10.1016/B978-0-12-336156-1.50061-6
Downloads
Published
How to Cite
Issue
Section
