Anatomy-VLM: A Fine-grained Vision-Language Model for Medical Interpretation

Anatomy-VLM represents a significant advancement in medical imaging interpretation, addressing the complexities of disease diagnosis that arise from imaging heterogeneity. Traditional vision-language models often overlook critical details necessary for accurate assessments. By mimicking the human-centric approach of clinicians, Anatomy-VLM effectively localizes anatomical features and enriches them with structured knowledge, leading to expert-level clinical interpretations. Its validation on both in- and out-of-distribution datasets demonstrates its robust capabilities, making it a valuable tool for image segmentation tasks and enhancing the overall diagnostic process in radiology.