EchoPrime, a novel AI model trained on a vast dataset of echocardiogram videos and reports, leverages multi-view video analysis and vision-language integration to achieve state-of-the-art performance in comprehensive echocardiography interpretation, surpassing existing task-specific models and foundation models.
EchoPrime 是一種基於視覺語言的深度學習模型,透過整合多個超聲心動圖視圖的資訊,實現對超聲心動圖的全面自動解讀,並在多項心臟結構和功能評估指標上達到甚至超越人類專家的表現。