Exploring Zero-Shot Anomaly Detection with CLIP in Medical Imaging: Are We There Yet?
This work addresses the problem of limited data annotation in medical imaging for clinicians, but it is incremental as it highlights the need for further adaptation of existing models.
The paper evaluated CLIP-based models for zero-shot anomaly detection in brain tumor detection using the BraTS-MET dataset, finding that while they show promise in transferring general knowledge, their performance falls short of clinical precision requirements.
Zero-shot anomaly detection (ZSAD) offers potential for identifying anomalies in medical imaging without task-specific training. In this paper, we evaluate CLIP-based models, originally developed for industrial tasks, on brain tumor detection using the BraTS-MET dataset. Our analysis examines their ability to detect medical-specific anomalies with no or minimal supervision, addressing the challenges posed by limited data annotation. While these models show promise in transferring general knowledge to medical tasks, their performance falls short of the precision required for clinical use. Our findings highlight the need for further adaptation before CLIP-based models can be reliably applied to medical anomaly detection.