Comparing GAN, Diffusion, and Diffusion-GAN for Single-Image Deraining of UAV Imagery

Salsabilah Aulia Rahman; Laksmita Rahadianti

doi:10.30595/juita.v14i1.28994

Authors

Salsabilah Aulia Rahman Universitas Indonesia
Laksmita Rahadianti Universitas Indonesia

DOI:

https://doi.org/10.30595/juita.v14i1.28994

Keywords:

Single-image deraining, UAV imagery, generative adversarial networks, diffusion models, diffusion–GAN.

Abstract

Single-image deraining for Unmanned Aerial Vehicle (UAV) imagery remains challenging due to non-uniform rain patterns, motion blur, and real-time processing requirements. Existing generative paradigms, including Generative Adversarial Networks (GAN), Diffusion, and Diffusion–GAN, each face inherent trade-offs among restoration quality, stability, and efficiency. To address the lack of unified and fair benchmarking across these paradigms, this study presents a systematic and controlled comparative evaluation of three representative models, including TBGAN, WeatherDiff, and SupResDiffGAN, to assess their relative performance in UAV deraining tasks. The models are evaluated on the UAV-Rain1K and Rain100L datasets using PSNR, SSIM, and inference efficiency metrics to support informed selection of paradigms for UAV applications. Experimental results show that WeatherDiff achieves the highest fidelity with 19.99 dB PSNR, 0.8375 SSIM on UAV-Rain1K and 29.51 dB PSNR, 0.9093 SSIM on Rain100L. TBGAN yields sharper details but lower structural consistency, whereas SupResDiffGAN offers balanced performance with 19.03 dB PSNR and 0.7053 SSIM on UAV-Rain1K and 28.51 dB PSNR and 0.8681 SSIM on Rain100L, with faster inference. These findings highlight the practical trade-offs among the three paradigms and demonstrate that diffusion–GAN frameworks provide the most practical solution for UAV deraining, combining diffusion stability with adversarial sharpness for real-time restoration.

Author Biographies

Salsabilah Aulia Rahman, Universitas Indonesia

Faculty of Computer Science

Laksmita Rahadianti, Universitas Indonesia

Faculty of Computer Science

References

[1] Y. Li, Y. Monno, and M. Okutomi, “Single Image Deraining Network with Rain Embedding Consistency and Layered LSTM,” in Proc. IEEE/CVF Winter Conf. on Applications of Computer Vision (WACV), 2022, pp. 3957–3966. doi: 10.1109/WACV51458.2022.00401.

[2] L. Zhao, J. Long, and T. Zhong, “A Deep Learning-Based Two-Branch Generative Adversarial Network for Image De-Raining,” Sensors, vol. 24, no. 20, p. 6724, 2024, doi: 10.3390/s24206724.

[3] P. Li, J. Jin, G. Jin, L. Fan, X. Gao, T. Song, and X. Chen, “Deep Scale-space Mining Network for Single Image Deraining,” in Proc. IEEE/CVF Conf. on Computer Vision and Pattern Recognition Workshops (CVPRW), 2022, pp. 4275–4284. doi: 10.1109/CVPRW56347.2022.00473.

[4] Y. Tomida, T. Katayama, T. Song, and T. Shimamoto, “Efficient Deraining Model Using Transformer and Kernel Basis Attention for UAVs,” in Proc. 2024 Int. Tech. Conf. on Circuits/Systems, Computers, and Communications (ITC-CSCC), 2024, pp. 1–5. doi: 10.1109/ITC-CSCC62988.2024.10628256.

[5] W. Li, C. Li, H. Jiang, Y. Wang, S. Wu, and Z. Wu, “High-Resolution Aerial Image Restoration with Latent Diffusion Models,” in Proc. IEEE Int. Conf. on Unmanned Systems (ICUS), 2024, pp. 1874–1878. doi: 10.1109/ICUS61736.2024.10840082.

[6] Q. Yi, J. Li, Q. Dai, F. Fang, G. Zhang, and T. Zeng, “Structure-Preserving Deraining with Residue Channel Prior Guidance,” presented at the 2021 IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 4218–4227.

[7] M. H. Kolekar, S. Bose, and A. Pai, “SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal,” IEEE Access, vol. 12, 2024, doi: 10.1109/ACCESS.2024.3375909.

[8] Y. Wang, Z. Ma, and H. Su, “Research on Hole Feature Matching Based on Image Restoration,” in Proc. Int. Conf. on Computers, Information Processing and Advanced Education (CIPAE), 2022, pp. 385–389. doi: 10.1109/CIPAE55637.2022.00087.

[9] O. Özdenizci and R. Legenstein, “Restoring Vision in Adverse Weather Conditions with Patch-Based Denoising Diffusion Models,” vol. 45, 2023, doi: 10.1109/TPAMI.2023.3238179.

[10] A. F. Akbar, P. D. W. Ayu, and D. P. Hostiadi, “Performance Analysis of Deep Learning Architectures in Classifying Fake and Real Images,” JUITA J. Inform. 132 167–176, vol. 13, no. 2, 2025, doi: https://doi.org/10.30595/juita.v13i2.25790.

[11] I. J. Goodfellow, “Generative Adversarial Networks,” Adv. Neural Inf. Process. Syst., vol. 3, 2014, doi: 10.1145/3422622.

[12] J. Ho, A. Jain, and P. Abbeel, “Denoising Diffusion Probabilistic Models,” Adv Neural Inf Process Sys, vol. 33, pp. 6840–6851, 2020, doi: https://doi.org/10.48550/arXiv.2006.11239.

[13] Z. Wang, H. Zheng, P. He, W. Chen, and M. Zhou, “Diffusion-GAN: Training GANs with Diffusion,” ArXiv Prepr. ArXiv220602262, 2022, doi: https://doi.org/10.48550/arXiv.2206.02262.

[14] P. Salehi, A. Chalechale, and M. Taghizadeh, “Generative Adversarial Networks (GANs): An Overview of Theoretical Model, Evaluation Metrics, and Recent Developments,” ArXiv Prepr. ArXiv200513178, 2020, doi: 10.48550/arXiv.2005.13178.

[15] H. Cho, H.-K. Shin, Y. Jang, S.-J. Ko, and S.-W. Jung, “PD-CR: Patch-Based Diffusion Using Constrained Refinement for Image Restoration,” IEEE Signal Process. Lett., vol. 31, 2024, doi: 10.1109/LSP.2024.3381908.

[16] X. Lin, J. He, Z. Chen, Z. Lyu, B. Dai, F. Yu, W. Ouyang, Y. Qiao, and C. Dong, “DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior,” ArXiv Prepr. ArXiv230815070, 2024, doi: https://doi.org/10.48550/arXiv.2308.15070.

[17] L. T. Trinh and T. Hamagami, “Latent Denoising Diffusion GAN: Faster Sampling, Higher Image Quality,” IEEE Access, vol. 12, pp. 78161–78172, 2024, doi: 10.1109/ACCESS.2024.3406535.

[18] D. Kopeć, W. Kozłowski, M. Wizerkaniuk, D. Krutul, J. Kocoń, and M. Zięba, “SupResDiffGAN: A New Approach for the Super-Resolution Task,” ArXiv Prepr. ArXiv250413622, 2025, doi: https://doi.org/10.48550/arXiv.2504.13622.

[19] W. Chang, H. Chen, X. He, X. Chen, and L. Shen, “UAV-Rain1k: A Benchmark for Raindrop Removal from UAV Aerial Imagery,” ArXiv Prepr. ArXiv240205773, 2024, doi: https://doi.org/10.48550/arXiv.2402.05773.

[20] W. Simoes and M. D. Sá, “PSNR and SSIM: Evaluation of the Imperceptibility Quality of Images Transmitted over Wireless Networks,” Procedia Comput. Sci., vol. 251, pp. 463–470, 2024, doi: 10.1016/j.procs.2024.11.134.

[21] J. Nilsson and T. Akenine-Möller, “Understanding SSIM.” 2020. doi: https://doi.org/10.48550/arXiv.2006.13846.