Utilizing Segment Anything Model For Assessing Localization of GRAD-CAM in Medical Imaging
This work addresses the need for rigorous testing of AI interpretability in medical imaging, though it is incremental as it builds on existing saliency map methods.
The paper tackled the problem of evaluating saliency map localization in medical imaging by proposing the use of the Segment Anything Model (SAM) to improve accuracy and eliminate the need for human annotations, achieving high similarity to existing metrics.
The introduction of saliency map algorithms as an approach for assessing the interoperability of images has allowed for a deeper understanding of current black-box models with Artificial Intelligence. Their rise in popularity has led to these algorithms being applied in multiple fields, including medical imaging. With a classification task as important as those in the medical domain, a need for rigorous testing of their capabilities arises. Current works examine capabilities through assessing the localization of saliency maps upon medical abnormalities within an image, through comparisons with human annotations. We propose utilizing Segment Anything Model (SAM) to both further the accuracy of such existing metrics, while also generalizing beyond the need for human annotations. Our results show both high degrees of similarity to existing metrics while also highlighting the capabilities of this methodology to beyond human-annotation. Furthermore, we explore the applications (and challenges) of SAM within the medical domain, including image pre-processing before segmenting, natural language proposals to SAM in the form of CLIP-SAM, and SAM accuracy across multiple medical imaging datasets.