Segmentor-Guided Counterfactual Fine-Tuning for Locally Coherent and Targeted Image Synthesis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xia, Tian, Sinclair, Matthew, Schuh, Andreas, Ribeiro, Fabio De Sousa, Mehta, Raghav, Rasal, Rajat, Puyol-Antón, Esther, Gerber, Samuel, Petersen, Kersten, Schaap, Michiel, Glocker, Ben
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916984625561600
author Xia, Tian
Sinclair, Matthew
Schuh, Andreas
Ribeiro, Fabio De Sousa
Mehta, Raghav
Rasal, Rajat
Puyol-Antón, Esther
Gerber, Samuel
Petersen, Kersten
Schaap, Michiel
Glocker, Ben
author_facet Xia, Tian
Sinclair, Matthew
Schuh, Andreas
Ribeiro, Fabio De Sousa
Mehta, Raghav
Rasal, Rajat
Puyol-Antón, Esther
Gerber, Samuel
Petersen, Kersten
Schaap, Michiel
Glocker, Ben
contents Counterfactual image generation is a powerful tool for augmenting training data, de-biasing datasets, and modeling disease. Current approaches rely on external classifiers or regressors to increase the effectiveness of subject-level interventions (e.g., changing the patient's age). For structure-specific interventions (e.g., changing the area of the left lung in a chest radiograph), we show that this is insufficient, and can result in undesirable global effects across the image domain. Previous work used pixel-level label maps as guidance, requiring a user to provide hypothetical segmentations which are tedious and difficult to obtain. We propose Segmentor-guided Counterfactual Fine-Tuning (Seg-CFT), which preserves the simplicity of intervening on scalar-valued, structure-specific variables while producing locally coherent and effective counterfactuals. We demonstrate the capability of generating realistic chest radiographs, and we show promising results for modeling coronary artery disease. Code: https://github.com/biomedia-mira/seg-cft.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24913
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Segmentor-Guided Counterfactual Fine-Tuning for Locally Coherent and Targeted Image Synthesis
Xia, Tian
Sinclair, Matthew
Schuh, Andreas
Ribeiro, Fabio De Sousa
Mehta, Raghav
Rasal, Rajat
Puyol-Antón, Esther
Gerber, Samuel
Petersen, Kersten
Schaap, Michiel
Glocker, Ben
Computer Vision and Pattern Recognition
Artificial Intelligence
Counterfactual image generation is a powerful tool for augmenting training data, de-biasing datasets, and modeling disease. Current approaches rely on external classifiers or regressors to increase the effectiveness of subject-level interventions (e.g., changing the patient's age). For structure-specific interventions (e.g., changing the area of the left lung in a chest radiograph), we show that this is insufficient, and can result in undesirable global effects across the image domain. Previous work used pixel-level label maps as guidance, requiring a user to provide hypothetical segmentations which are tedious and difficult to obtain. We propose Segmentor-guided Counterfactual Fine-Tuning (Seg-CFT), which preserves the simplicity of intervening on scalar-valued, structure-specific variables while producing locally coherent and effective counterfactuals. We demonstrate the capability of generating realistic chest radiographs, and we show promising results for modeling coronary artery disease. Code: https://github.com/biomedia-mira/seg-cft.
title Segmentor-Guided Counterfactual Fine-Tuning for Locally Coherent and Targeted Image Synthesis
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2509.24913