RFMSR: Residual Flow Matching for Image Super-Resolution
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: AI Radio. Generated commentary on the latest Artificial Intelligence papers.
Tom: We're looking at the paper: "RFMSR: Residual Flow Matching for Image Super-Resolution".
Jane: Here's the gist of the paper. Image super-resolution (ISR) has witnessed remarkable progress with diffusion models and flow matching<ref:2607.12753#pg2> RFMSR, a vision-only framework that centers the source distribution at the LQ latent to reduce transport distance and preserve structural priors throughout the flow trajectory<ref:2607.12753#pg5> The proposed method achieves comparable or even superior perceptual quality compared to state-of-the-art (SOTA) methods<ref:2607.12753#pg9> Pruning redundant reasoning tokens The paper introduces Residual Flow Matching for Image SuperResolution (RFMSR), a vision-only framework that reformulates flow matching for the ISR task<ref:2607.12753#pg5> The key insight is that the LQ image provides substantial prior knowledge about the target HQ image, and the flow should leverage this information rather than starting from pure noise<ref:2607.12753#pg9> Specifically, it defines the source distribution as a Gaussian centered at the LQ latent y0 with variance σ squared I<ref:2607.12753#pg9> This residual flow formulation reduces the transport distance and preserves structural priors throughout the trajectory, enabling more efficient and faithful reconstruction<ref:2607.12753#pg9> How the training signal is built The authors propose a two-phase training strategy to achieve high-quality single-step generation while fully retaining multi-step inference capability<ref:2607.12753#pg9> Phase I involves training the model using only the CFM objective defined in Eq.
Tom: (7), which establishes a smooth transport from the source distribution N (y0, σ squared I) to the target HQ latent distribution<ref:2607.12753#pg9> Phase II applies end-to-end supervision to the single-step prediction quality while concurrently retaining the standard velocity loss of flow matching across all timesteps<ref:2607.12753#pg9> The full Phase II objective is LPhase II = L(θ)+λLPIPS LLPIPS(ˆx0, x0)+λGAN Ladv(ˆx0, x0)<ref:2607.12753#pg9> Where the gains come from The effectiveness of the residual flow formulation is demonstrated by comparing it against standard rectified flow, which consistently outperforms it across all metrics and step counts<ref:2607.12753#pg9> The two-phase training strategy maintains strong single-step fidelity on par with Phase I, and can further improve perceptual quality by increasing inference steps<ref:2607.12753#pg9> Ablation studies confirm the effectiveness of the residual flow formulation and the two-phase training design<ref:2607.12753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff, maintaining strong single-step quality while fully retaining multi-step inference capability<ref:2607.12753#pg9> Comparison of Flow Matching Strategies The comparison of flow matching strategies validates the gains brought by the residual flow formulation when compared against standard rectified flow under identical training settings<ref:2607.12753#pg9> The residual flow consistently outperforms the rectified flow across all metrics and step counts, in terms of both distortion fidelity (e.g., PSNR, LPIPS) and perceptual quality (e.g., MUSIQ, CLIPIQA)<ref:2607.12753#pg9> The consistent improvement can be attributed to centering the source distribution at y0, which shortens the transport path and preserves structural priors<ref:2607.12753#pg9> Ablation Study on Noise Level Setting σ = 0.0 reduces the residual flow to a deterministic L2 regression, which achieves comparable fidelity metrics (PSNR and SSIM) but suffers from severe perceptual degradation across all perceptual indicators<ref:2607.12753#pg9> Introducing a modest noise level (σ = 0.5) dramatically improves perceptual quality (e.g.
Jane: MUSIQ, CLIPIQA), with substantial gains in LPIPS and NIQE<ref:2607.12753#pg9> Further increasing σ to 1.0 or 2.0 yields additional gains in reconstruction fidelity (PSNR, SSIM) while maintaining good perceptual quality<ref:2607.12753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff, maintaining strong single-step quality while fully retaining multi-step inference capability<ref:2607.12753#pg9> Additional Visual Comparisons RFMSR consistently produces results with the most natural texture and fewest artifacts among multi-step methods<ref:2607.12753#pg9> In the single-step setting, RFMSR demonstrates clear advantages over competing methods: VOSR produces plausible but slightly blurred outputs, while InvSR and OSEDiff frequently introduce local artifacts and distorted structures<ref:2607.12753#pg9> Notably, RFMSR maintains accurate glyph reconstruction for text content and preserves fine-grained facial textures with high fidelity<ref:2607.12753#pg9> The residual flow formulation combined with two-phase training provides a robust foundation for both high-quality single-step generation and progressive multistep refinement<ref:2607.12753#pg9> Efficiency Comparison Table 7 presents the parameter counts and inference runtime of the model components of each method<ref:2607.12753#pg9> Since RFMSR does not employ the classifier-free guidance (CFG) inference used in VOSR, its runtime is slightly shorter<ref:2607.12753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff<ref:2607.12753#pg9> The consistent improvement can be attributed to centering the source distribution at y0, which shortens the transport path and preserves structural priors<ref:2607.12753#pg9> Limitation Our model architecture is designed based on the 0.5B-parameter variant of VOSR<ref:2607.12753#pg9> Further scaling up the model capacity and training from scratch on larger datasets may yield additional performance improvements<ref:2607.12753#pg9> The current framework is limited to the ×4 ISR setting and has not been validated on broader degradation types<ref:2607.12753#pg9> Furthermore, incorporating more advanced end-to-end reconstruction losses, such as larger GAN discriminators or better perceptual losses, may further boost single-step generation quality<ref:2607.12753#pg9> In future work, we will further investigate these issues<ref:2607.12753#pg9> References
1: Eirikur Agustsson and Radu Timofte.
Tom: Ntire 2017 challenge on single image super-resolution: Dataset and study<ref:2607.12753#pg9>
2: Jianrui Cai, Hui Zeng, Hongwei Yong, Zisheng Cao, and Lei Zhang.
Paper summary: Tom: So we're looking at this paper called "RFMSR: Residual Flow Matching for Image SuperResolution," and it seems like they're tackling the problem of image super-resolution using diffusion models and flow matching.
Jane: That sounds really technical, Tom. What's the main idea here? What are they trying to claim is different from what we usually see in this area?
Tom: Well, the core thesis is that standard flow formulations just transport from pure noise to the data distribution, which means they throw away structural information already present in the low-quality image.
Lu: Exactly. The authors propose a vision-only framework called RFMSR that centers the source distribution at the low-quality latent image, which is what they call y0, to reduce that transport distance and keep those structural priors preserved throughout the flow trajectory.
Meng: So they're saying it leverages knowledge from the input LQ image instead of just starting from pure noise? That sounds like a big idea for efficiency.
Lalam: I can confirm, Lalam says their key insight is that the low-quality image provides substantial prior knowledge about the high-quality target image, and the flow should use that information instead of starting from pure noise.
Tom: Right, so they define the source distribution as a Gaussian centered at y0 with a variance sigma squared I. That residual flow formulation reduces transport distance and preserves those structural priors throughout the whole trajectory, which makes reconstruction more efficient and faithful.
Jane: But if we want to actually use this in the real world, it needs to work well for single-step generation too, right? Because standard methods often sacrifice that multi-step capability when they try to speed things up.
Tom: That's where their two-phase training strategy comes in. Phase one trains the model using only the CFM objective defined in equation seven, which establishes a smooth transport from that source distribution N (y0, sigma squared I) to the target high-quality latent distribution HQ.
Lu: And phase two is where they apply end-to-end supervision for single-step prediction quality while keeping the standard velocity loss of flow matching across all timesteps. That's how they keep that multi-step inference ability intact, even when aiming for a fast, one-step result.
Paper summary: Meng: From an engineering standpoint, that two-phase approach sounds smart because it separates learning the general flow from fine-tuning the final output quality. What does that look like in terms of what we're actually training?
Tom: Phase two involves an objective called LPhase II which combines the CFM velocity loss with LPIPS, which measures perceptual quality, and a GAN adversarial loss to check if the predicted image looks realistic.
Jane: So they aren't just focusing on pixel accuracy, but they're also making sure it looks good to the human eye through those perceptual losses. That’s important for super-resolution work.
Lalam: I can tell that by combining LPhase II, they get the best overall tradeoff, maintaining strong single-step quality while fully retaining that multi-step inference capability for arbitrary step counts >
Tom: The comparison against standard rectified flow shows it consistently outperforms it across all metrics and step counts in terms of both distortion fidelity like PSNR and LPIPS, as well as perceptual quality like MUSIQ and CLIPIQA.
Lu: That consistent improvement is directly attributed to centering the source distribution at y0, which shortens the transport path and preserves those structural priors throughout the flow trajectory >
Meng: But what about when we look at their experiments regarding noise level? Does tweaking that variance change things much?
Tom: They did an ablation study on noise level. Setting sigma equals zero reduces the residual flow to a deterministic L2 regression, which achieves comparable fidelity metrics like PSNR and SSIM, but it suffers from severe perceptual degradation across all perceptual indicators.
Jane: Wait, so removing the noise term makes it look technically similar but way worse visually? That's a key detail for us to understand.
Lu: Introducing a modest noise level, like sigma equals zero point five, dramatically improves perceptual quality on things like MUSIQ and CLIPIQA, with substantial gains in LPIPS and NIQE <ref:2607.12753#pg1>. And increasing sigma up to one point zero or two point zero yields additional gains in reconstruction fidelity like PSNR and SSIM while still maintaining good perceptual quality.
Paper summary: Tom: So the full Phase two, combining the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff for them when balancing single-step quality and multi-step inference capability >
Meng: Looking at efficiency, they have a table showing parameter counts and inference runtime for each method. Since RFMSR doesn't use classifier-free guidance like VOSR does, its runtime is slightly shorter.
Jane: So for someone who's actually running this on a device, the inference time difference is small but real? What kind of limitations are they acknowledging about this work?
Tom: The paper points out a few things. First, their model architecture is based on the zero point five billion parameter variant of VOSR. They suggest that further scaling up the model capacity and training from scratch on larger datasets could bring additional performance improvements.
Lu: Also, they admit the current framework is limited to the times four image super-resolution setting and hasn't been validated on broader degradation types yet. And they note that incorporating more advanced end-to-end reconstruction losses, like larger GAN discriminators or better perceptual losses, could further boost single-step generation quality.
Meng: That makes sense for production. If you're deploying this right now, you have to know those limits regarding the degradation types and where it stops working >
Jane: So basically, they’ve built a very robust foundation that works well for both high-quality single-step generation and progressive multi-step refinement, but we still have room to improve the model size and test it on more real-world scenarios.
Tom: The paper "RFMSR: Residual Flow Matching for Image SuperResolution" is focused on how to leverage the existing low-quality image structure in flow matching to get better results without losing control over the inference steps >
Lu: It’s a significant refinement because it solves that problem of discarding structural information when moving from noise, which is what standard flow formulations often do >
Meng: For production, this means we have a solid baseline that handles single-step needs well while keeping the full multi-step capability available for refinement >
Jane: It’s interesting how they used the concept of residual shifting from ResShift to inspire their approach here >
Conclusion: Tom: So we've seen how they use this residual flow idea and how it works with a two-phase training setup for single step and multi step results.
Jane: It sounds like the authors are trying to make sure you get that high quality output even when you just want one quick result, without losing the ability to refine it later.
Lu: Exactly, they’re balancing those two needs using that phase two objective that includes things like LPIPS and GAN loss.
Meng: From an engineering standpoint, the point is that this method keeps its structural priors better than some other flow methods we’ve seen before.
Lalam: I’m confident their approach means we get a really solid foundation for realistic image reconstruction.
Tom: So the authors called it RFMSR, Residual Flow Matching for Image Super-Resolution, and they did that by using this residual flow formulation instead of starting from pure noise.
Jane: It’s important to see how centering the source distribution at y zero helps them keep those important structural details intact during the entire image creation process.
Lu: That centering shortens the transport path, which means less distortion accumulates along the way as you go from low resolution to high resolution.
Meng: For me, what this means is that if we’re deploying models for real-world use, this residual flow looks like a way to get more reliable results without needing super long inference times compared to some other methods.
Tom: So it boils down to using the existing image structure as a guide for the generation process rather than just treating it like random noise and trying to build something from scratch.
Jane: And they’ve shown that this approach beats standard rectified flow on all the metrics we looked at, both in terms of accuracy and how natural the resulting images look.
Lu: It really suggests that focusing on preserving those initial image priors through a guided flow path is a very effective strategy for diffusion models in this area.
Meng: It’s interesting that they’re scaling this work to see if it can handle different noise levels too, which is something I think would matter when we test it on actual messy data.
Tom: Yeah, so while the results are strong across the board, there are still things they point out about scaling up the model size and testing it on totally different degradation types that we should watch for in future work.
Huazhong University of Science and Technology
cs.CV
Submitted: 2026-07-14
Updated: 2026-11-13
Code: https://github.com/FazeHsw/RFMSR
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 92/100
The gist: Image super-resolution (ISR) has witnessed remarkable progress with diffusion models and flow matching<ref:2607.12753#pg2> RFMSR…
Key concepts
- Residual Flow Matching
- This formulation redefines flow matching by setting the source distribution as a Gaussian centered at the low-quality latent (LQ) y0. This residual approach minimizes the transport distance during the flow trajectory, ensuring that structural priors from the LQ image are preserved throughout the reconstruction process, making it more efficient.
- Two-Phase Training Strategy
- The training uses a two-stage process. Phase I trains the model using only a standard CFM objective to establish smooth transport. Phase II then applies end-to-end supervision for single-step prediction quality while simultaneously keeping the standard flow matching velocity loss across all timesteps, balancing fidelity and quality.
- Source Distribution Centering
- The core idea is centering the source distribution at y0 rather than starting from pure noise. This modification significantly shortens the required transport path for image reconstruction. By leveraging the inherent structural information present in the LQ latent, this centering helps preserve important structural priors, resulting in better reconstructions.
- Phase II Objective
- The full Phase II objective combines several losses: L(θ) for prediction quality, LPIPS to enforce perceptual similarity between generated and ground truth images, and a GAN loss (Ladv) to enhance visual realism. This comprehensive loss function is designed to achieve the best trade-off between strong single-step fidelity and high perceptual quality.
Terminology
Summary
Image super-resolution (ISR) has witnessed remarkable progress with diffusion models and flow matching<ref:2607. p. 012753#pg2> RFMSR, a vision-only framework that centers the source distribution at the LQ latent to reduce transport distance and preserve structural priors throughout the flow trajectory<ref:2607. p. 012753#pg5> The proposed method achieves comparable or even superior perceptual quality compared to state-of-the-art (SOTA) methods<ref:2607. p. 012753#pg9>
Pruning redundant reasoning tokens
The paper introduces Residual Flow Matching for Image SuperResolution (RFMSR), a vision-only framework that reformulates flow matching for the ISR task<ref:2607.12753#pg5> The key insight is that the LQ image provides substantial prior knowledge about the target HQ image, and the flow should leverage this information rather than starting from pure noise<ref:2607. p. 112753#pg9> Specifically, it defines the source distribution as a Gaussian centered at the LQ latent y0 with variance σ squared I<ref:2607. p. 112753#pg9> This residual flow formulation reduces the transport distance and preserves structural priors throughout the trajectory, enabling more efficient and faithful reconstruction<ref:2607. p. 112753#pg9>
How the training signal is built
The authors propose a two-phase training strategy to achieve high-quality single-step generation while fully retaining multi-step inference capability<ref:2607.12753#pg9> Phase I involves training the model using only the CFM objective defined in Eq. p. 3 (7), which establishes a smooth transport from the source distribution N (y0, σ squared I) to the target HQ latent distribution<ref:2607. p. 312753#pg9> Phase II applies end-to-end supervision to the single-step prediction quality while concurrently retaining the standard velocity loss of flow matching across all timesteps<ref:2607. p. 112753#pg9> The full Phase II objective is LPhase II = L(θ)+λLPIPS LLPIPS(ˆx0, x0)+λGAN Ladv(ˆx0, x0)<ref:2607. p. 312753#pg9>
Where the gains come from
The effectiveness of the residual flow formulation is demonstrated by comparing it against standard rectified flow, which consistently outperforms it across all metrics and step counts<ref:2607.12753#pg9> The two-phase training strategy maintains strong single-step fidelity on par with Phase I, and can further improve perceptual quality by increasing inference steps<ref:2607. p. 712753#pg9> Ablation studies confirm the effectiveness of the residual flow formulation and the two-phase training design<ref:2607. p. 712753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff, maintaining strong single-step quality while fully retaining multi-step inference capability<ref:2607. p. 7 p. 712753#pg9>
Comparison of Flow Matching Strategies
The comparison of flow matching strategies validates the gains brought by the residual flow formulation when compared against standard rectified flow under identical training settings<ref:2607.12753#pg9> The residual flow consistently outperforms the rectified flow across all metrics and step counts, in terms of both distortion fidelity (e. p. 6g., PSNR, LPIPS) and perceptual quality (e. p. 4g., MUSIQ, CLIPIQA)<ref:2607. p. 412753#pg9> The consistent improvement can be attributed to centering the source distribution at y0, which shortens the transport path and preserves structural priors<ref:2607. p. 7 p. 712753#pg9>
Ablation Study on Noise Level
Setting σ = 0.0 reduces the residual flow to a deterministic L2 regression, which achieves comparable fidelity metrics (PSNR and SSIM) but suffers from severe perceptual degradation across all perceptual indicators<ref:2607. p. 312753#pg9> Introducing a modest noise level (σ = 0. p. 35) dramatically improves perceptual quality (e. p. 0g. MUSIQ, CLIPIQA), with substantial gains in LPIPS and NIQE<ref:2607. p. 412753#pg9> Further increasing σ to 1. p. 30 or 2.0 yields additional gains in reconstruction fidelity (PSNR, SSIM) while maintaining good perceptual quality<ref:2607. p. 712753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff, maintaining strong single-step quality while fully retaining multi-step inference capability<ref:2607.12753#pg9>
Additional Visual Comparisons
RFMSR consistently produces results with the most natural texture and fewest artifacts among multi-step methods<ref:2607.12753#pg9> In the single-step setting, RFMSR demonstrates clear advantages over competing methods: VOSR produces plausible but slightly blurred outputs, while InvSR and OSEDiff frequently introduce local artifacts and distorted structures<ref:2607. p. 512753#pg9> Notably, RFMSR maintains accurate glyph reconstruction for text content and preserves fine-grained facial textures with high fidelity<ref:2607. p. 512753#pg9> The residual flow formulation combined with two-phase training provides a robust foundation for both high-quality single-step generation and progressive multistep refinement<ref:2607. p. 512753#pg9>
Efficiency Comparison
Table 7 presents the parameter counts and inference runtime of the model components of each method<ref:2607.12753#pg9> Since RFMSR does not employ the classifier-free guidance (CFG) inference used in VOSR, its runtime is slightly shorter<ref:2607. p. 812753#pg9> The full Phase II, which combines the CFM velocity loss with LPIPS and GAN, achieves the best overall tradeoff<ref:2607. p. 712753#pg9> The consistent improvement can be attributed to centering the source distribution at y0, which shortens the transport path and preserves structural priors<ref:2607.12753#pg9>
Limitation
Our model architecture is designed based on the 0.5B-parameter variant of VOSR<ref:2607. p. 412753#pg9> Further scaling up the model capacity and training from scratch on larger datasets may yield additional performance improvements<ref:2607. p. 712753#pg9> The current framework is limited to the ×4 ISR setting and has not been validated on broader degradation types<ref:2607. p. 712753#pg9> Furthermore, incorporating more advanced end-to-end reconstruction losses, such as larger GAN discriminators or better perceptual losses, may further boost single-step generation quality<ref:2607. p. 712753#pg9> In future work, we will further investigate these issues<ref:2607. p. 712753#pg9>
References
[1] Eirikur Agustsson and Radu Timofte. Ntire 2017 challenge on single image super-resolution: Dataset and study<ref:2607. p. 812753#pg9>
[2] Jianrui Cai, Hui Zeng, Hongwei Yong, Zisheng Cao, and Lei Zhang. Toward real-world single image super-resolution: A new benchmark and a new model<ref:2607. p. 812753#pg9>
[3] Xiangyu Chen, Xintao Wang, Jiantao Zhou, Yu Qiao, and Chao Dong. Activating more pixels in image superresolution transformer<ref:2607. p. 812753#pg9>
[4] Keyan Ding, Kede Ma, Shiqi Wang, and Eero P Simoncelli. Image quality assessment: Unifying structure and texture similarity<ref:2607. p. 812753#pg9>
[5] Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Learning a deep convolutional network for image super-resolution<ref:2607. p. 812753#pg9>
[6] Linwei Dong, Qingnan Fan, Yihong Guo, Zhonghao Wang, Qi Zhang, Jinwei Chen, Yawei Luo, and Changqing Zou. Tsd-sr: One-step diffusion with target score distillation for real-world image super-resolution<ref:2607. p. 812753#pg9>
[7] Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas Muller, Harry Saini, Yam Levi, Dominik ¨ Lorenz, Axel Sauer, Frederic Boesel, et al. Scaling rectified flow transformers for high-resolution image synthesis<ref:2607. p. 812753#pg9>
[8] Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium<ref:2607. p. 812753#pg9>
[9] Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance<ref:2607. p. 812753#pg9>
[10] Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models<ref:2607. p. 812753#pg9>
[11] Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. Image-to-image translation with conditional adversarial networks<ref:2607. p. 812753#pg9>
[12] Junjie Ke, Qifei Wang, Yilin Wang, Peyman Milanfar, and Feng Yang. Musiq: Multi-scale image quality transformer<ref:2607. p. 812753#pg9>
[13] Christian Ledig, Lucas Theis, Ferenc Huszar, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, et al. Photorealistic single image super-resolution using a generative adversarial network<ref:2607. p. 812753#pg9>
[14] Jianze Li, Jiezhang Cao, Yong Guo, Wenbo Li, and Yulun Zhang. One diffusion step to real-world super-resolution via flow trajectory distillation<ref:2607. p. 812753#pg9>
[15] Yawei Li, Kai Zhang, Jingyun Liang, Jiezhang Cao, Ce Liu, Rui Gong, Yulun Zhang, Hao Tang, Yun Liu, Denis Demandolx, et al. Lsdir: A large scale dataset for image restoration<ref:2607. p. 812753#pg9>
[16] Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration using swin transformer<ref:2607. p. 812753#pg9>
[17] Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee. Enhanced deep residual networks for single image super-resolution<ref:2607. p. 812753#pg9>
[18] Xinqi Lin, Jingwen He, Ziyan Chen, Zhaoyang Lyu, Bo Dai, Fanghua Yu, Yu Qiao, Wanli Ouyang, and Chao Dong. Diffbir: Toward blind image restoration with generative diffusion prior<ref:2607. p. 812753#pg9>
[19] Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling<ref:2607. p. 012753#pg9>
[20] Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow<ref:2607. p. 812753#pg9>
[21] Xingchao Liu, Xiwen Zhang, Jianzhu Ma, Jian Peng, et al. Instaflow: One step is enough for high-quality diffusionbased text-to-image generation<ref:2607. p. 812753#pg9>
[22] Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization<ref:2607. p. 812753#pg9>
[23] Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Making a “completely blind” image quality analyzer<ref:2607. p. 812753#pg9>
[24] Takeru Miyato, Toshiki Kataoka, Masanori Koyama, and Yuichi Yoshida. Spectral normalization for generative adversarial networks<ref:2607. p. 812753#pg9>
[25] Mehdi Noroozi, Isma Hadji, Brais Martinez, Adrian Bulat, and Georgios Tzimiropoulos. You only need one step: Fast super-resolution with stable diffusion via scale distillation<ref:2607. p. 812753#pg9>
[26] Maxime Oquab, Timothee Darcet, Th ´ eo Moutakanni, Huy ´ Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision<ref:2607. p. 912753#pg9>
[27] Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Muller, Joe Penna, and ¨ Robin Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis<ref:2607. p. 912753#pg9>
[28] Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bjorn Ommer. High-resolution image ¨ synthesis with latent diffusion models<ref:2607. p. 912753#pg9>
[29] Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge<ref:2607. p. 912753#pg9>
[30] Chitwan Saharia, Jonathan Ho, William Chan, Tim Salimans, David J Fleet, and Mohammad Norouzi. Image superresolution via iterative refinement<ref:2607. p. 912753#pg9>
[31] Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models<ref:2607. p. 912753#pg9>
[32] Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations<ref:2607. p. 912753#pg9>
[33] Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency models<ref:2607. p. 012753#pg9>
[34] Stability.ai. Stable diffusion<ref:2607. p. 012753#pg9>
[35] Jianyi Wang, Kelvin CK Chan, and Chen Change Loy. Exploring clip for assessing the look and feel of images<ref:2607. p. 912753#pg9>
[36] Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin CK Chan, and Chen Change Loy. Exploiting diffusion prior for real-world image super-resolution<ref:2607. p. 912753#pg9>
[37] Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. Esrgan: Enhanced super-resolution generative adversarial networks<ref:2607. p. 912753#pg9>
[38] Xintao Wang, Liangbin Xie, Chao Dong, and Ying Shan. Real-esrgan: Training real-world blind super-resolution with pure synthetic data<ref:2607. p. 912753#pg9>
[39] Yufei Wang, Wenhan Yang, Xinyuan Chen, Yaohui Wang, Lanqing Guo, Lap-Pui Chau, Ziwei Liu, Yu Qiao, Alex C Kot, and Bihan Wen. Sinsr: diffusion-based image superresolution in a single step<ref:2607. p. 912753#pg9>
[40] Z Wang, C Lu, Y Wang, F Bao, C Li, H Su, and J Zhu. Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation<ref:2607. p. 912753#pg9>
[41] Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity<ref:2607. p. 912753#pg9>
[42] Rongyuan Wu, Lingchen Sun, Zhiyuan Ma, and Lei Zhang. One-step effective diffusion network for real-world image super-resolution<ref:2607. p. 912753#pg9>
[43] Rongyuan Wu, Tao Yang, Lingchen Sun, Zhengqiang Zhang, Shuai Li, and Lei Zhang. Seesr: Towards semanticsaware real-world image super-resolution<ref:2607. p. 912753#pg9>
[44] Rongyuan Wu, Lingchen Sun, Zhengqiang Zhang, Xiangtao Kong, Jixin Zhao, Shihao Wang, and Lei Zhang. Vosr: A vision-only generative model for image super-resolution<ref:2607. p. 912753#pg9>
[45] Jiaqi Xu, Wenbo Li, Haoze Sun, Fan Li, Zhixin Wang, Long Peng, Jingjing Ren, Haoran Yang, Xiaowei Hu, Renjing Pei, et al. Fast image super-resolution via consistency rectified flow<ref:2607. p. 912753#pg9>
[46] Sidi Yang, Tianhe Wu, Shuwei Shi, Shanshan Lao, Yuan Gong, Mingdeng Cao, Jiahao Wang, and Yujiu Yang. Maniqa: Multi-dimension attention network for no-reference image quality assessment<ref:2607. p. 912753#pg9>
[47] Tao Yang, Rongyuan Wu, Peiran Ren, Xuansong Xie, and Lei Zhang. Pixel-aware stable diffusion for realistic image super-resolution and personalized stylization<ref:2607. p. 912753#pg9>
[48] Jingfeng Yao, Bin Yang, and Xinggang Wang. Reconstruction vs. p. 1 generation: Taming optimization dilemma in latent diffusion models<ref:2607. p. 912753#pg9>
[49] Fanghua Yu, Jinjin Gu, Zheyuan Li, Jinfan Hu, Xiangtao Kong, Xintao Wang, Jingwen He, Yu Qiao, and Chao Dong. Scaling up to excellence: Practicing model scaling for photorealistic image restoration in the wild<ref:2607. p. 912753#pg9>
[50] Zongsheng Yue, Jianyi Wang, and Chen Change Loy. Resshift: Efficient diffusion model for image superresolution by residual shifting<ref:2607. p. 912753#pg9>
[51] Zongsheng Yue, Kang Liao, and Chen Change Loy. Arbitrary-steps image super-resolution via diffusion inversion<ref:2607. p. 312753#pg9>
[52] Kai Zhang, Jingyun Liang, Luc Van Gool, and Radu Timofte. Designing a practical degradation model for deep blind image super-resolution<ref:2607. p. 012753#pg9>
[53] Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric<ref:2607. p. 0
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Residual Flow Matching for ISR allows vision-only frameworks to leverage structural information from low-quality (LQ) inputs by defining a source distribution centered at the LQ latent y0, which
reduces transport distance and preserves structural priors throughout the flow trajectory.
This enables more efficient and faithful reconstruction compared to standard flow matching starting from pure noise. -
The two-phase training strategy allows for high-quality single-step generation while
retaining the standard velocity loss of flow matching across all timesteps,
whichendows the model with flexible inference ranging from a single forward pass to multi-step refinement.
This means the system can perform both fast one-step generation and quality-enhanced multi-step refinement. -
The system can be optimized for perceptual quality by tuning the noise level sigma; setting sigma to 1.0 provides
a favorable trade-off between distortion fidelity and perceptual naturalness across all datasets,
leading to results thatconsistently matches or exceeds SeeSR and VOSR in perceptual quality while maintaining reasonable distortion fidelity.
Sources
- Classifier-Free Diffusion Guidance
- One Diffusion Step to Real-World Super-Resolution via Flow Trajectory Distillation
- Flow Matching for Generative Modeling
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- Decoupled Weight Decay Regularization
- Spectral Normalization for Generative Adversarial Networks
- DINOv2: Learning Robust Visual Features without Supervision
- Denoising Diffusion Implicit Models
- Score-Based Generative Modeling through Stochastic Differential Equations
- Consistency Models
- ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs