English

Prior-Guided Residual Diffusion: Calibrated and Efficient Medical Image Segmentation

Computer Vision and Pattern Recognition 2026-02-03 v2

Abstract

Ambiguity in medical image segmentation calls for models that capture full conditional distributions rather than a single point estimate. We present Prior-Guided Residual Diffusion (PGRD), a diffusion-based framework that learns voxel-wise distributions while maintaining strong calibration and practical sampling efficiency. PGRD embeds discrete labels as one-hot targets in a continuous space to align segmentation with diffusion modeling. A coarse prior predictor provides step-wise guidance; the diffusion network then learns the residual to the prior, accelerating convergence and improving calibration. A deep diffusion supervision scheme further stabilizes training by supervising intermediate time steps. Evaluated on representative MRI and CT datasets, PGRD achieves higher Dice scores and lower NLL/ECE values than Bayesian, ensemble, Probabilistic U-Net, and vanilla diffusion baselines, while requiring fewer sampling steps to reach strong performance.

Keywords

Cite

@article{arxiv.2509.01330,
  title  = {Prior-Guided Residual Diffusion: Calibrated and Efficient Medical Image Segmentation},
  author = {Fuyou Mao and Beining Wu and Yanfeng Jiang and Han Xue and Yan Tang and Hao Zhang},
  journal= {arXiv preprint arXiv:2509.01330},
  year   = {2026}
}

Comments

Withdrawn by the authors to conduct further methodological refinement and address concerns regarding the originality of the current implementation

R2 v1 2026-07-01T05:15:05.853Z