Accelerating Diffusion Models for Inverse Problems through Shortcut Sampling
This work addresses the computational inefficiency of diffusion models in inverse problems for applications like image restoration, though it is incremental as it builds on existing zero-shot methods.
The paper tackles the problem of accelerating diffusion models for solving inverse problems by proposing Shortcut Sampling for Diffusion (SSD), which uses a transitional state to bridge measurement and restored images, achieving competitive results with only 30 NFEs compared to state-of-the-art methods requiring 100 NFEs.
Diffusion models have recently demonstrated an impressive ability to address inverse problems in an unsupervised manner. While existing methods primarily focus on modifying the posterior sampling process, the potential of the forward process remains largely unexplored. In this work, we propose Shortcut Sampling for Diffusion(SSD), a novel approach for solving inverse problems in a zero-shot manner. Instead of initiating from random noise, the core concept of SSD is to find a specific transitional state that bridges the measurement image y and the restored image x. By utilizing the shortcut path of "input - transitional state - output", SSD can achieve precise restoration with fewer steps. To derive the transitional state during the forward process, we introduce Distortion Adaptive Inversion. Moreover, we apply back projection as additional consistency constraints during the generation process. Experimentally, we demonstrate SSD's effectiveness on multiple representative IR tasks. Our method achieves competitive results with only 30 NFEs compared to state-of-the-art zero-shot methods(100 NFEs) and outperforms them with 100 NFEs in certain tasks. Code is available at https://github.com/GongyeLiu/SSD