Saved in:
Bibliographic Details
Main Authors: Sun, Lingchen, Wu, Rongyuan, Ma, Zhiyuan, Liu, Shuaizheng, Yi, Qiaosi, Zhang, Lei
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.03017
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912306220236800
author Sun, Lingchen
Wu, Rongyuan
Ma, Zhiyuan
Liu, Shuaizheng
Yi, Qiaosi
Zhang, Lei
author_facet Sun, Lingchen
Wu, Rongyuan
Ma, Zhiyuan
Liu, Shuaizheng
Yi, Qiaosi
Zhang, Lei
contents Diffusion prior-based methods have shown impressive results in real-world image super-resolution (SR). However, most existing methods entangle pixel-level and semantic-level SR objectives in the training process, struggling to balance pixel-wise fidelity and perceptual quality. Meanwhile, users have varying preferences on SR results, thus it is demanded to develop an adjustable SR model that can be tailored to different fidelity-perception preferences during inference without re-training. We present Pixel-level and Semantic-level Adjustable SR (PiSA-SR), which learns two LoRA modules upon the pre-trained stable-diffusion (SD) model to achieve improved and adjustable SR results. We first formulate the SD-based SR problem as learning the residual between the low-quality input and the high-quality output, then show that the learning objective can be decoupled into two distinct LoRA weight spaces: one is characterized by the $\ell_2$-loss for pixel-level regression, and another is characterized by the LPIPS and classifier score distillation losses to extract semantic information from pre-trained classification and SD models. In its default setting, PiSA-SR can be performed in a single diffusion step, achieving leading real-world SR results in both quality and efficiency. By introducing two adjustable guidance scales on the two LoRA modules to control the strengths of pixel-wise fidelity and semantic-level details during inference, PiSASR can offer flexible SR results according to user preference without re-training. Codes and models can be found at https://github.com/csslc/PiSA-SR.
format Preprint
id arxiv_https___arxiv_org_abs_2412_03017
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA Approach
Sun, Lingchen
Wu, Rongyuan
Ma, Zhiyuan
Liu, Shuaizheng
Yi, Qiaosi
Zhang, Lei
Computer Vision and Pattern Recognition
Diffusion prior-based methods have shown impressive results in real-world image super-resolution (SR). However, most existing methods entangle pixel-level and semantic-level SR objectives in the training process, struggling to balance pixel-wise fidelity and perceptual quality. Meanwhile, users have varying preferences on SR results, thus it is demanded to develop an adjustable SR model that can be tailored to different fidelity-perception preferences during inference without re-training. We present Pixel-level and Semantic-level Adjustable SR (PiSA-SR), which learns two LoRA modules upon the pre-trained stable-diffusion (SD) model to achieve improved and adjustable SR results. We first formulate the SD-based SR problem as learning the residual between the low-quality input and the high-quality output, then show that the learning objective can be decoupled into two distinct LoRA weight spaces: one is characterized by the $\ell_2$-loss for pixel-level regression, and another is characterized by the LPIPS and classifier score distillation losses to extract semantic information from pre-trained classification and SD models. In its default setting, PiSA-SR can be performed in a single diffusion step, achieving leading real-world SR results in both quality and efficiency. By introducing two adjustable guidance scales on the two LoRA modules to control the strengths of pixel-wise fidelity and semantic-level details during inference, PiSASR can offer flexible SR results according to user preference without re-training. Codes and models can be found at https://github.com/csslc/PiSA-SR.
title Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA Approach
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.03017