Consistent Diffusion: Denoising Diffusion Model with Data-Consistent Training for Image Restoration
This addresses image quality issues in restoration tasks for applications like photography or medical imaging, representing an incremental advance by refining training methods.
The paper tackles shape and color distortions in denoising diffusion models for image restoration by proposing data-consistent training, which accesses images with accumulated errors during training to correct them, resulting in significant improvements over state-of-the-art methods across five tasks.
In this work, we address the limitations of denoising diffusion models (DDMs) in image restoration tasks, particularly the shape and color distortions that can compromise image quality. While DDMs have demonstrated a promising performance in many applications such as text-to-image synthesis, their effectiveness in image restoration is often hindered by shape and color distortions. We observe that these issues arise from inconsistencies between the training and testing data used by DDMs. Based on our observation, we propose a novel training method, named data-consistent training, which allows the DDMs to access images with accumulated errors during training, thereby ensuring the model to learn to correct these errors. Experimental results show that, across five image restoration tasks, our method has significant improvements over state-of-the-art methods while effectively minimizing distortions and preserving image fidelity.