Generating healthy counterfactuals with denoising diffusion bridge models

2025-10-15

AI Summary

Generating healthy counterfactual pathology images while simultaneously removing lesions and preserving subject-specific anatomical features remains challenging. Method: This paper proposes a bidirectional generative framework based on the Denoising Diffusion Bridge Model (DDBM), which uniquely treats healthy images as the starting point and synthetic pathological images as the endpoint. By incorporating bidirectional structural priors, it jointly models the healthy-to-pathological and pathological-to-healthy state transitions within both forward and reverse diffusion processes. Contribution/Results: The method enables precise lesion editing while faithfully retaining individual anatomical structures, yielding high-fidelity reconstructions of healthy images. Experiments demonstrate that our approach significantly outperforms existing diffusion-based and fully supervised methods in segmentation and anomaly detection tasks. Generated images exhibit higher fidelity to real healthy scans, leading to measurable improvements in downstream medical analysis performance.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Generating healthy counterfactuals from pathological images holds significant promise in medical imaging, e.g., in anomaly detection or for application of analysis tools that are designed for healthy scans. These counterfactuals should represent what a patient's scan would plausibly look like in the absence of pathology, preserving individual anatomical characteristics while modifying only the pathological regions. Denoising diffusion probabilistic models (DDPMs) have become popular methods for generating healthy counterfactuals of pathology data. Typically, this involves training on solely healthy data with the assumption that a partial denoising process will be unable to model disease regions and will instead reconstruct a closely matched healthy counterpart. More recent methods have incorporated synthetic pathological images to better guide the diffusion process. However, it remains challenging to guide the generative process in a way that effectively balances the removal of anomalies with the retention of subject-specific features. To solve this problem, we propose a novel application of denoising diffusion bridge models (DDBMs) - which, unlike DDPMs, condition the diffusion process not only on the initial point (i.e., the healthy image), but also on the final point (i.e., a corresponding synthetically generated pathological image). Treating the pathological image as a structurally informative prior enables us to generate counterfactuals that closely match the patient's anatomy while selectively removing pathology. The results show that our DDBM outperforms previously proposed diffusion models and fully supervised approaches at segmentation and anomaly detection tasks.

Problem

• Generating healthy counterfactuals from pathological medical images
• Balancing anomaly removal with patient-specific feature retention
• Improving segmentation and anomaly detection in medical imaging

Innovation

• DDBMs condition diffusion on both healthy and pathological images
• Generating counterfactuals by selectively removing pathology
• Using pathological images as structurally informative priors