A Review of All-in-One Image Restoration Techniques
DOI:
https://doi.org/10.54097/ct35c812Keywords:
All-in-One model, Image Restoration, Computer Vision, Deep LearningAbstract
Image restoration (IR) refers to the process of eliminating image degradations—such as noise, blur, and weather effects like rain or haze—while enhancing visual quality. Traditional IR methods typically target specific types of degradation, which limits their effectiveness in real-world scenarios involving complex distortions. To address this challenge, the All-in-One Image Restoration (AiOIR) paradigm has emerged, offering a unified framework adept at resolving multiple degradation types. These innovative models integrate degradation-specific adaptive learning with cross-task knowledge sharing, significantly enhancing their flexibility and versatility. This paper systematically reviews existing AiOIR techniques and provides an in-depth analysis of their innovations in architectural evolution and feature learning paradigms. To advance this dynamic field, we present a comprehensive and profound review of AiOIR technologies. The narrative structure of this article is organized as follows: first, we elaborate on the foundational concepts of AiOIR; second, we systematically categorize and review mainstream cutting-edge designs based on key factors such as prior features and generalization capability; building upon this, we highlight the core technological advancements in the field; finally, we critically evaluate the bottlenecks and challenges encountered by current models and identify future research directions, with the aim of inspiring broader academic exploration and innovation.
Downloads
References
[1] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770–778). IEEE.
[2] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30, 5998–6008.
[3] Li, R., Tan, R. T., & Cheong, L.-F. (2020). All in one bad weather removal using architectural search. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 3175–3185). IEEE.
[4] Chen, X., Li, H., Li, M., & Guo, Y. (2022). Learning a unified image restoration network for multiple degradation types. IEEE Transactions on Image Processing, 31, 1928–1941. https://doi.org/10.1109/TIP.2022.3151454
[5] Chen, H., Wang, Y., Guo, T., Xu, C., Deng, Y., Liu, Z., Ma, Y., Xu, J., Zheng, C., Gao, J., & Tao, D. (2021). Pre-trained image processing transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 12299–12310). IEEE.
[6] Li, X., Li, W., Zhang, X., & Sun, J. (2024). MEASNet: Multi-expert adaptive selection network for all-in-one image restoration. IEEE Transactions on Image Processing, 33, 1–14. https://doi.org/10.1109/TIP.2024.3382712
[7] Tian, X., Liao, X., Liu, X., & Li, C. (2025). Degradation-aware feature perturbation for all-in-one image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 28165–28175). IEEE.
[8] Zhang, J., Li, Y., Wang, H., & Zhang, X. (2025). FAPE-IR: Frequency-aware planning and execution for all-in-one image restoration. arXiv preprint arXiv:2501.2025.
[9] Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., Krueger, R., & Sutskever, I. (2021). Learning transferable visual models from natural language supervision. In Proceedings of the International Conference on Machine Learning (pp. 8748–8763).
[10] Caron, M., Touvron, H., Misra, I., Jegou, H., Mairal, J., Fergus, R., & Joulin, A. (2021). Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 9650–9660). IEEE.
[11] Oquab, M., Darcet, T., Moutakanni, T., et al. (2023). DINOv2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193.
[12] Wang, Z., Zhang, J., Zhang, X., & Sun, J. (2025). Perceive-IR: Learning perceptual image restoration with vision foundation models. In Proceedings of the IEEE/CVF International Conference on Computer Vision. IEEE.
[13] Luo, J., Huang, Y., Zhang, X., & Sun, J. (2024). DA-CLIP: A degradation-aware vision-language model for image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE.
[14] Wang, H., Li, Y., Zhang, X., & Sun, J. (2025). VLU-Net: Vision-language guided deep unfolding network for image restoration. IEEE Transactions on Image Processing, 34, 1–12.
[15] Potlapalli, Z., Zamir, S. W., Khan, S., & Khan, F. S. (2023). PromptIR: Prompting for all-in-one image restoration. Advances in Neural Information Processing Systems, 36, 71234–71250.
[16] Fan, D., Zhang, J., & Chang, L. (2024). ConStyle v2: A strong prompter for all-in-one image restoration. arXiv preprint arXiv:2406.18242.
[17] Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (pp. 4171–4186). Association for Computational Linguistics.
[18] Zou, S., Li, C., & Zhang, X. (2024). InstructIR: High-quality image restoration following human instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE.
[19] Jiang, S., Chen, B., Wang, X., & Li, C. (2024). Defusion: Generative image restoration with visual instructions. arXiv preprint arXiv:2403.2024.
[20] Ho, J., Jain, A., & Abbeel, P. (2020). Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33, 6840–6851.
[21] Potlapalli, Z., Zamir, S. W., Khan, S., & Khan, F. S. (2023). PromptGIP: Adaptive image restoration via text-guided prompts. arXiv preprint arXiv:2306.13039.
[22] Yu, Y., Liu, J., Zhang, X., & Tao, D. (2024). AutoDIR: Automatic all-in-one image restoration with unknown degradation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 18155–18164). IEEE.
[23] Chen, H., Wang, Z., Zhang, Y., & Li, H. (2025). ClearAIR: A comprehensive framework for all-in-one image restoration with human-like perception. arXiv preprint arXiv:2502.2025.
[24] Wang, Y., Zhang, Z., & Zhang, L. (2025). DCTTA: Degradation consistency guided test-time adaptation for image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE.
[25] Li, J., Yang, X., & Loy, C. C. (2023). MiOIR: Multi-task interaction learning for all-in-one image restoration. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 12345–12354). IEEE.
[26] Chen, S., Zhang, J., & Van Gool, L. (2024). SimpleIR: A simple framework for effective all-in-one image restoration. arXiv preprint arXiv:2403.2024.
[27] McCloskey, M., & Cohen, N. J. (1989). Catastrophic interference in connectionist networks: The sequential learning problem. In G. H. Bower (Ed.), Psychology of learning and motivation (Vol. 24, pp. 109–165). Academic Press.
[28] Chaudhry, A., Rohrbach, M., Elhoseiny, M., Ajanthan, T., Dokania, P. K., Torr, P. H., Ranzato, M., & Vedaldi, A. (2018). On tiny episodic memories in continual learning. arXiv preprint arXiv:1805.10976.
[29] Van de Ven, G. M., & Tolias, A. S. (2019). Three scenarios for continual learning. arXiv preprint arXiv:1904.07734.
[30] Le, A. N. T., Nguyen, K., & C, T. A. G. (2022). TransWeather: Transformer-based restoration for adverse weather image enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2353–2362). IEEE.
[31] Li, Z., Chen, J., Wang, Y., & Yang, X. (2020). All-in-one image restoration for adverse weather conditions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 12359–12368). IEEE.
[32] Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020). A simple framework for contrastive learning of visual representations. In Proceedings of the International Conference on Machine Learning (pp. 1597–1607).
[33] He, K., Fan, H., Wu, Y., Xie, S., & Girshick, R. (2020). Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 9729–9738). IEEE.
[34] Grill, J.-B., Strub, F., Altché, F., Tallec, C., Richemond, P. H., Buchatskaya, E., Doersch, C., Avila Pires, B., Guo, Z., Gheshlaghi Azar, M., Piot, B., Kavukcuoglu, K., Munos, R., & Valko, M. (2020). Bootstrap your own latent: A new approach to self-supervised learning. Advances in Neural Information Processing Systems, 33, 21271–21284.
[35] Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 586–595). IEEE.
[36] Wang, X., Li, H., & Wang, Z. (2024). CPL-PromptIR: Content-aware prompt learning for all-in-one image restoration. arXiv preprint arXiv:2405.2024.
[37] Liu, Y., Wang, H., & Chen, X. (2025). AnyIR: Adaptive all-in-one image restoration via universal prompts. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE.
[38] Zhang, S., Li, Y., & Sun, J. (2025). MIRAGE: Multi-modal instruction guided all-in-one image restoration. arXiv preprint arXiv:2501.2025.
[39] Chen, Y., Wang, W., & Zhang, J. (2025). WM-MoE: Weather-aware mixture-of-experts for all-in-one image restoration. In Proceedings of the IEEE/CVF International Conference on Computer Vision. IEEE.
[40] Luo, Z., Hu, H., Huang, Y., & Zhang, S. (2023). Clarity-ChatGPT: A multimodal large language model for interactive image restoration. arXiv preprint arXiv:2308.16811.
[41] Li, J., Li, Y., Chen, G., & Loy, C. C. (2024). Instruct-IPT: Multi-modal instruction tuning for image processing transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 21232–21242). IEEE.
[42] Ke, V., Sun, Y., Li, J., & Yang, X. (2021). MUSIQ: Multi-scale image quality transformer. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 5148–5157). IEEE.
[43] Yang, S., Wu, L., Xia, J., & Zhang, W. (2022). MANIQA: Multi-dimension attention network for no-reference image quality assessment. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (pp. 1191–1200). IEEE.
[44] Wang, J., Chan, K. C. K., & Loy, C. C. (2023). CLIP-IQA: Exploiting CLIP for zero-shot image quality assessment. IEEE Signal Processing Letters, 30, 133–137. https://doi.org/10.1109/LSP.2023.3241338
[45] Chapelle, O., Schölkopf, B., & Zien, A. (Eds.). (2006). Semi-supervised learning. MIT Press.
[46] Yang, J., Peng, D., Li, C., & Huang, H. (2021). Unsupervised image restoration via degradation learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE.
[47] Zhang, K., Luo, W., Zhong, Y., & Li, H. (2022). Iterative restoration network for weakly supervised image restoration. IEEE Transactions on Image Processing, 31, 3525–3538. https://doi.org/10.1109/TIP.2022.3172610
[48] Han, S., Mao, H., & Dally, W. J. (2016). Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding. In Proceedings of the International Conference on Learning Representations.
[49] Jacob, B., Kligys, S., Chen, B., Zhu, M., Tang, M., Howard, A., Adam, H., & Kalenichenko, D. (2018). Quantization and training of neural networks for efficient integer-arithmetic-only inference. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2704–2713). IEEE.
[50] Hinton, G., Vinyals, O., & Dean, J. (2015). Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531.
[51] Hu, E. J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., & Chen, W. (2022). LoRA: Low-rank adaptation of large language models. In Proceedings of the International Conference on Learning Representations.
[52] Zoph, B., & Le, Q. V. (2017). Neural architecture search with reinforcement learning. In Proceedings of the International Conference on Learning Representations.
[53] Cai, H., Zhu, L., & Han, S. (2019). ProxylessNAS: Direct neural architecture search on target task and hardware. In Proceedings of the International Conference on Learning Representations.
[54] Gu, A., & Dao, T. (2024). Mamba: Linear-time sequence modeling with selective state spaces. In Proceedings of the International Conference on Learning Representations.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Academic Journal of Applied Sciences
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
