MedCondDiff: Lightweight, Robust, Semantically Guided Diffusion for Medical Image Segmentation

TL;DR

MedCondDiff is a lightweight, robust diffusion-based framework for multi-organ medical image segmentation that leverages semantic priors for improved performance and efficiency across various imaging modalities.

Contribution

Introduces MedCondDiff, a semantically guided diffusion model using a Pyramid Vision Transformer to enhance medical image segmentation efficiency and robustness.

Findings

Achieves competitive segmentation performance across multiple organs and modalities.

Reduces inference time and VRAM usage compared to traditional diffusion models.

Demonstrates robustness and efficiency in diverse medical imaging tasks.

Abstract

We introduce MedCondDiff, a diffusion-based framework for multi-organ medical image segmentation that is efficient and anatomically grounded. The model conditions the denoising process on semantic priors extracted by a Pyramid Vision Transformer (PVT) backbone, yielding a semantically guided and lightweight diffusion architecture. This design improves robustness while reducing both inference time and VRAM usage compared to conventional diffusion models. Experiments on multi-organ, multi-modality datasets demonstrate that MedCondDiff delivers competitive performance across anatomical regions and imaging modalities, underscoring the potential of semantically guided diffusion models as an effective class of architectures for medical imaging tasks.