CheXplaining in Style: Counterfactual Explanations for Chest X-rays using StyleGAN
This work addresses reliability concerns for clinicians using AI in medical diagnostics by providing interpretable explanations, though it is incremental as it builds on existing StyleGAN and counterfactual explanation techniques.
The paper tackled the problem of understanding black-box deep learning models in medical image analysis by generating counterfactual explanations for chest X-rays, using a StyleGAN-based method (StyleEx) and EigenFind to reduce computation time, with clinical evaluation by radiologists showing relevancy.
Deep learning models used in medical image analysis are prone to raising reliability concerns due to their black-box nature. To shed light on these black-box models, previous works predominantly focus on identifying the contribution of input features to the diagnosis, i.e., feature attribution. In this work, we explore counterfactual explanations to identify what patterns the models rely on for diagnosis. Specifically, we investigate the effect of changing features within chest X-rays on the classifier's output to understand its decision mechanism. We leverage a StyleGAN-based approach (StyleEx) to create counterfactual explanations for chest X-rays by manipulating specific latent directions in their latent space. In addition, we propose EigenFind to significantly reduce the computation time of generated explanations. We clinically evaluate the relevancy of our counterfactual explanations with the help of radiologists. Our code is publicly available.