Adaptive Anomaly detection: Application to brain metastases

This project explores the use of visual language models for anomaly detection in medical applications. Our goal is to create a tool capable of zero-shot anomaly detection, enabling it to identify anomalies without prior training on specific cases. As a case study, we focus on brain MRI scans to detect brain metastases. By leveraging state-of-the-art visual and textual integration, this research aims to provide a groundbreaking solution for early diagnosis and precision medicine in oncology.