Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Kim, Bryan Sangwoo, Kim, Jeongsol, Ye, Jong Chul
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914462112415744
author Kim, Bryan Sangwoo
Kim, Jeongsol
Ye, Jong Chul
author_facet Kim, Bryan Sangwoo
Kim, Jeongsol
Ye, Jong Chul
contents Modern single-image super-resolution (SISR) models deliver photo-realistic results at the scale factors on which they are trained, but collapse when asked to magnify far beyond that regime. We address this scalability bottleneck with Chain-of-Zoom (CoZ), a model-agnostic framework that factorizes SISR into an autoregressive chain of intermediate scale-states with multi-scale-aware prompts. CoZ repeatedly re-uses a backbone SR model, decomposing the conditional probability into tractable sub-problems to achieve extreme resolutions without additional training. Because visual cues diminish at high magnifications, we augment each zoom step with multi-scale-aware text prompts generated by a vision-language model (VLM). The prompt extractor itself is fine-tuned using Generalized Reward Policy Optimization (GRPO) with a critic VLM, aligning text guidance towards human preference. Experiments show that a standard 4x diffusion SR model wrapped in CoZ attains beyond 256x enlargement with high perceptual quality and fidelity. Project Page: https://bryanswkim.github.io/chain-of-zoom/.
format Preprint
id arxiv_https___arxiv_org_abs_2505_18600
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
Kim, Bryan Sangwoo
Kim, Jeongsol
Ye, Jong Chul
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Modern single-image super-resolution (SISR) models deliver photo-realistic results at the scale factors on which they are trained, but collapse when asked to magnify far beyond that regime. We address this scalability bottleneck with Chain-of-Zoom (CoZ), a model-agnostic framework that factorizes SISR into an autoregressive chain of intermediate scale-states with multi-scale-aware prompts. CoZ repeatedly re-uses a backbone SR model, decomposing the conditional probability into tractable sub-problems to achieve extreme resolutions without additional training. Because visual cues diminish at high magnifications, we augment each zoom step with multi-scale-aware text prompts generated by a vision-language model (VLM). The prompt extractor itself is fine-tuned using Generalized Reward Policy Optimization (GRPO) with a critic VLM, aligning text guidance towards human preference. Experiments show that a standard 4x diffusion SR model wrapped in CoZ attains beyond 256x enlargement with high perceptual quality and fidelity. Project Page: https://bryanswkim.github.io/chain-of-zoom/.
title Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2505.18600