Efficient and Training-Free Single-Image Diffusion Models

Researchers have developed a training-free diffusion model that generates images matching a single reference image's internal structure by using a patch-based approach with a closed-form denoiser, eliminating the need for computationally expensive neural network training. The method achieves state-of-the-art quality while being significantly faster, enabling megapixel generation in one second and gigapixel generation in minutes, with applications including image stylization, symmetrization, and retargeting.