CTSR: Controllable Fidelity-Realness Trade-off Distillation for Real-World Image Super Resolution

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Li, Runyi, Chen, Bin, Zhang, Jian, Timofte, Radu
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912282018054144
author Li, Runyi
Chen, Bin
Zhang, Jian
Timofte, Radu
author_facet Li, Runyi
Chen, Bin
Zhang, Jian
Timofte, Radu
contents Real-world image super-resolution is a critical image processing task, where two key evaluation criteria are the fidelity to the original image and the visual realness of the generated results. Although existing methods based on diffusion models excel in visual realness by leveraging strong priors, they often struggle to achieve an effective balance between fidelity and realness. In our preliminary experiments, we observe that a linear combination of multiple models outperforms individual models, motivating us to harness the strengths of different models for a more effective trade-off. Based on this insight, we propose a distillation-based approach that leverages the geometric decomposition of both fidelity and realness, alongside the performance advantages of multiple teacher models, to strike a more balanced trade-off. Furthermore, we explore the controllability of this trade-off, enabling a flexible and adjustable super-resolution process, which we call CTSR (Controllable Trade-off Super-Resolution). Experiments conducted on several real-world image super-resolution benchmarks demonstrate that our method surpasses existing state-of-the-art approaches, achieving superior performance across both fidelity and realness metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2503_14272
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CTSR: Controllable Fidelity-Realness Trade-off Distillation for Real-World Image Super Resolution
Li, Runyi
Chen, Bin
Zhang, Jian
Timofte, Radu
Computer Vision and Pattern Recognition
Image and Video Processing
Real-world image super-resolution is a critical image processing task, where two key evaluation criteria are the fidelity to the original image and the visual realness of the generated results. Although existing methods based on diffusion models excel in visual realness by leveraging strong priors, they often struggle to achieve an effective balance between fidelity and realness. In our preliminary experiments, we observe that a linear combination of multiple models outperforms individual models, motivating us to harness the strengths of different models for a more effective trade-off. Based on this insight, we propose a distillation-based approach that leverages the geometric decomposition of both fidelity and realness, alongside the performance advantages of multiple teacher models, to strike a more balanced trade-off. Furthermore, we explore the controllability of this trade-off, enabling a flexible and adjustable super-resolution process, which we call CTSR (Controllable Trade-off Super-Resolution). Experiments conducted on several real-world image super-resolution benchmarks demonstrate that our method surpasses existing state-of-the-art approaches, achieving superior performance across both fidelity and realness metrics.
title CTSR: Controllable Fidelity-Realness Trade-off Distillation for Real-World Image Super Resolution
topic Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2503.14272