SOURCE-LINKED INTELLIGENCE

Unsupervised Anatomical Feature Learning via Diffusion Models: Enhanced Medical Image Segmentation with Denoising Diffusion Probabilistic Models

arXiv · AI, language, vision and robotics · article · Aug 26, 2026 · UTC

Acquiring pixel-level annotations for medical image segmentation is a severe bottleneck. Traditional U-Net architectures, while effective, learn local texture patterns and lack awareness of global anatomical structures, leading to boundary delineation failures in low-data regimes. This research paper proposes utilizing unsupervised Denoising Diffusion Probabilistic Models (DDPMs) to extract anatomical features. We train a DDPM on 21 unlabeled abdominal CT scans to learn structural representations, transferring the encoder weights to a downstream segmentation task evaluated on the BTCV multi-or

Read original source ↗ Open in workspace

recordType

paper

region

Global