MRD-UNet: Detail-Preserving Low-Light Image Enhancement with Multiscale Residual Dense Networks and Edge-Texture Guided Loss
Keywords:
Contextual attention module, Edge-texture guided loss, Interlevel residual learning, Low-light image enhancement, Multiscale residual dense networksAbstract
Low-light image enhancement is an important task in computer vision, particularly for night photography, autonomous driving, and video surveillance. Although recent deep learning methods perform well in reducing global noise, many models still degrade micro-structural details and texture fidelity because of excessive smoothing under extreme low-light conditions. This study proposes MRD-UNet, an encoder–decoder model designed to enhance low-light images while preserving structural integrity and textural sharpness. The model introduces two main components. First, the MRD-Block combines residual connections, dense connections, and multiscale convolutions to capture both global contextual features and fine local details. Second, an Edge-Texture Guided Loss based on Sobel and Laplacian operators is used to guide the learning process, improving edge consistency and high-frequency detail preservation. Experiments on the LOL-V1 benchmark show that MRD-UNet outperforms several state-of-the-art methods, achieving an SSIM of 0.911 and PSNR of 26.09 dB.
Downloads
References
Hao, S., Han, X., Guo, Y., Xu, X., and Wang, M. (2020). Low-light image enhancement with semi-decoupled decomposition. IEEE Transactions on Multimedia, 22(12), 3025–3038.
Zhang, Y., Guo, X., Ma, J., Liu, W., and Zhang, J. (2021). Beyond brightening low-light images. International Journal of Computer Vision, 129(4), 1013–1037.
Cheng, H. D., and Shi, X. J. (2004). A simple and effective histogram equalization approach to image enhancement. Digital Signal Processing, 14(2), 158–170.
Vijayalakshmi, D., Nath, M. K., and Acharya, O. P. (2020). A comprehensive survey on image contrast enhancement techniques in spatial domain. Sensing and Imaging, 21(1), 21-40.
Pizer, S. M. (1990). Contrast-limited adaptive histogram equalization: Speed and effectiveness. Proceedings of the First Conference on Visualization in Biomedical Computing, 1990, 337-345.
Jobson, D. J., Rahman, Z., and Woodell, G. A. (1997). Properties and performance of a center/surround retinex. IEEE Transactions on Image Processing, 6(3), 451–462.
Fu, X., Zeng, D., Huang, Y., Zhang, X.-P., and Ding, X. (2016). A weighted variational model for simultaneous reflectance and illumination estimation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, 2782-2790.
Wei, C., Wang, W., Yang, W., and Liu, J. (2018). Deep retinex decomposition for low-light enhancement. arXiv preprint arXiv, 1808, 04560.
Zhang, Y., Zhang, J., and Guo, X. (2019). Kindling the darkness: A practical low-light image enhancer. arXiv preprint arXiv, 1905, 04161.
Tao, L., Zhu, C., Xiang, G., Li, Y., Jia, H., and Xie, X. (2017). LLCNN: A convolutional neural network for low-light image enhancement. Proceedings of the IEEE Visual Communications and Image Processing (VCIP), 2017, 1-4.
Jiang, Y., Gong, X., Liu, D., Cheng, Y., Fang, C., Shen, X., Yang, J., Zhou, P., and Wang, Z. (2021). EnlightenGAN: Deep light enhancement without paired supervision. IEEE Transactions on Image Processing, 30, 2340-2349.
Xu, X., Wang, R., Fu, C.-W., and Jia, J. (2022). SNR-aware low-light image enhancement. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, 17714-17724.
Zamir, S.W., Arora, A., Khan, S.H., Hayat, M., Khan, F.S., and Yang, M. (2022). Restormer: Efficient transformer for high-resolution image restoration. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, 5718-5729.
Wang, T., Zhang, K., Shen, T., Luo, W., Stenger, B., and Lu, T. (2022). Ultra-high-definition low-light image enhancement: A benchmark and transformer-based method. arXiv preprint arXiv, 2212, 11548.
Ronneberger, O., Fischer, P., and Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation. arXiv preprint arXiv, 1505, 04597.
Cao, J., Chen, Z., Cui, H., Ji, X., Wang, X., Liang, Y., and Tian, Y. (2023). Improved wavelet prediction superresolution reconstruction based on U-Net. IET Image Processing, 17, 3464–3476.
He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, 770-778.
Huang, G., Liu, Z., and Weinberger, K. Q. (2017). Densely connected convolutional networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, 2261-2269.
Snell, J., Ridgeway, K., Liao, R., Roads, B. D., Mozer, M. C., and Zemel, R. S. (2015). Learning to generate images with perceptual similarity metrics. arXiv preprint arXiv, 1511, 06409.
Yang, W., Wang, W., Huang, H., Wang, S., and Liu, J. (2021). Sparse gradient regularized deep retinex network for robust low-light image enhancement. IEEE Transactions on Image Processing, 30, 2072–2086.
Surono, S., Rivaldi, M., Dewi, D. A., and Irsalinda, N. (2023). New approach to image segmentation: U-Net convolutional network for multiresolution CT image lung segmentation. Emerging Science Journal, 7(2), 498–506.
Zhang, Y., Tian, Y., Kong, Y., Zhong, B., and Fu, Y.R. (2021). Residual Dense Network for Image Restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(7), 2480-2495.
Wang, L., Zhao, L., Zhong, T., and Wu, C. (2024). Low-light image enhancement using generative adversarial networks. Scientific Reports, 14, 18489.
Yin, M., and Yang, J. (2025). ILR-Net: Low-light image enhancement network based on iterative learning mechanism and Retinex theory. PLoS ONE, 20(2), e0314541.
Lim, C. C., Loh, Y. P., and Wong, L.-K. (2023). LAU-Net: A low light image enhancer with attention and resizing mechanisms. Signal Processing: Image Communication, 115, 116971.
Jingchun, Z., Su, G. E., and Sunar, M. S. (2024). Low-light image enhancement: A comprehensive review on methods, datasets and evaluation metrics. Journal of King Saud University - Computer and Information Sciences, 36(10), 102234.
Jiang, B., Wang, X., Yang, N., Liu, Y., Chen, X., and Wu, Q. (2025). Semantic-aware low-light image enhancement by learning from multiple color spaces. Applied Sciences, 15(10), 5556.
Shen, X., Li, H., Li, Y., and Zhang, W. (2025). ColorBoost-LLIE: A multi-loss guided low-light image enhancement algorithm with decoupled color and luminance restoration. Displays, 87, 102979.
Avi-Aharon, M., Arbelle, A., and Raviv, T. R. (2020). DeepHist: Differentiable joint and color histogram layers for image-to-image translation. arXiv preprint arXiv, 2005, 03995.
Wang, Z., Simoncelli, E. P., and Bovik, A. C. (2003). Multiscale structural similarity for image quality assessment. Proceedings of the IEEE Asilomar Conference on Signals, Systems and Computers, 2, 1398–1402.
Marr, D., and Hildreth, E. (1980). Theory of edge detection. Proceedings of the Royal Society of London B, 207, 187–217.
Wang, Z., Bovik, A. C., Sheikh, H. R., and Simoncelli, E. P. (2004). Image quality assessment: From error visibility to structural similarity. IEEE Transactions on Image Processing, 13(4), 600–612.
Zhang, R., Isola, P., Efros, A.A., Shechtman, E., and Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, 586-595.
Mittal, A., Soundararajan, R., and Bovik, A. C. (2013). Making a completely blind image quality analyzer. IEEE Signal Processing Letters, 20(3), 209–212.
Simonyan, K., and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv, 1409, 1556.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
