Abstract: Denoising diffusion probabilistic models (DDPMs) exhibit significant potential in the realm of medical image segmentation. Nevertheless, current DDPM implementations rely on original image ...