SinSR: Diffusion-Based Image Super-Resolution in a Single Step
Authors:
Top Cited Papers
- 16 June 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 25796-25805
Abstract
While super-resolution (SR) methods based on diffusion models exhibit promising results, their practical application is hindered by the substantial number of required inference steps. Recent methods utilize the degraded images in the initial state, thereby shortening the Markov chain. Nevertheless, these solutions either rely on a precise formulation of the degradation process or still necessitate a relatively lengthy generation path (e.g., 15 iterations). To enhance inference speed, we propose a simple yet effective method for achieving single-step SR generation, named SinSR. Specifically, we first derive a deterministic sampling process from the most recent state-of-the-art (SOTA) method for accelerating diffusion-based SR. This allows the mapping between the input random noise and the generated high-resolution image to be obtained in a reduced and acceptable number of inference steps during training. We show that this deterministic mapping can be distilled into a student model that performs SR within only one inference step. Additionally, we propose a novel consistency-preserving loss to simultaneously leverage the ground-truth image during the distillation process, ensuring that the performance of the student model is not solely bound by the feature manifold of the teacher model, resulting in further performance improvement. Extensive experiments conducted on synthetic and real-world datasets demonstrate that the proposed method can achieve comparable or even superior performance compared to both previous SOTA methods and the teacher model, in just one sampling step, resulting in a remarkable up to × 10 speedup for inference. Our code will be released at https://github.com/wyf0912/SinSR/.Keywords
All Related Versions
This publication has 29 references indexed in Scilit:
- Efficient and Degradation-Adaptive Network for Real-World Image Super-Resolution2022
- Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems through Stochastic Contraction2022
- High-Resolution Image Synthesis with Latent Diffusion Models2022
- LAR-SR: A Local Autoregressive Model for Image Super-Resolution2022
- SRFlow: Learning the Super-Resolution Space with Normalizing Flow2020
- PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models2020
- The Unreasonable Effectiveness of Deep Features as a Perceptual Metric2018
- Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network2017
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks2016
- ImageNet: A large-scale hierarchical image database2009