RadEdit: Stress-Testing Biomedical Vision Models via Diffusion Image Editing
Abstract
Biomedical imaging datasets are often small and biased, meaning that real-world performance of predictive models can be substantially lower than expected from internal testing. This work proposes using generative image editing to simulate dataset shifts and diagnose failure modes of biomedical vision models; this can be used in advance of deployment to assess readiness, potentially reducing cost and patient harm. Existing editing methods can produce undesirable changes, with spurious correlations learned due to the co-occurrence of disease and treatment interventions, limiting practical applicability. To address this, we train a text-to-image diffusion model on multiple chest X-ray datasets and introduce a new editing method, , that uses multiple image masks, if present, to constrain changes and ensure consistency in the edited images, minimising bias. We consider three types of dataset shifts: acquisition shift, manifestation shift, and population shift, and demonstrate that our approach can diagnose failures and quantify model robustness without additional data collection, complementing more qualitative tools for explainable AI.
Cite
Text
Pérez-García et al. "RadEdit: Stress-Testing Biomedical Vision Models via Diffusion Image Editing." Proceedings of the European Conference on Computer Vision (ECCV), 2024. doi:10.1007/978-3-031-73254-6_21Markdown
[Pérez-García et al. "RadEdit: Stress-Testing Biomedical Vision Models via Diffusion Image Editing." Proceedings of the European Conference on Computer Vision (ECCV), 2024.](https://mlanthology.org/eccv/2024/perezgarcia2024eccv-radedit/) doi:10.1007/978-3-031-73254-6_21BibTeX
@inproceedings{perezgarcia2024eccv-radedit,
title = {{RadEdit: Stress-Testing Biomedical Vision Models via Diffusion Image Editing}},
author = {Pérez-García, Fernando and Bond-Taylor, Sam and Sanchez, Pedro and van Breugel, Boris and de Castro, Daniel Coelho and Sharma, Harshita and Salvatelli, Valentina and Wetscherek, Maria Teodora A and Richardson, Hannah CM and Matthew, Lungren and Nori, Aditya and Alvarez-Valle, Javier and Oktay, Ozan and Ilse, Maximilian},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2024},
doi = {10.1007/978-3-031-73254-6_21},
url = {https://mlanthology.org/eccv/2024/perezgarcia2024eccv-radedit/}
}