The HeaortaNet is developed by the TW-CVAI team. The HeaortaNet is a deep learning model trained by 70,000 axial images from 200 patients with verified annotations of the pericardiumaorta. It shortens the time for data processing from 60 minutes, by manual segmentation of both pericardiumaorta, to 0.4 seconds. The segmentation accuracy, as assessed by dice similarity coefficient, is 94.8 for the pericardium,91.6 for the aorta. The imaging-based Cardiovascular Risk Prediction module was constructed by analyzing data from the National Health Insurance Databank.

The HeaortaNet is a deep learning model based on the UNetattention gate. It had been trained by 70,000 axial images from 200 patients with verified annotations of the pericardiumaorta. The HeaortaNet has been listed in the NVIDIA GPU cloud (NGC) for AI research, the only one segmentation tool for pericardiumaorta in NGC worldwide. Both the quantitative analyses of cardiac/aortic calcificationepicardial adipose tissue modulethe imaging-based Cardiovascular Risk Prediction module are breakthroughs.

The value of this model could be analyzed in 3 sectors: medical, artificial intelligence (AI),health/insurance. In the medical field, it can enhance the diagnostic value of non-contrast chest CT, reduce the workload of radiologists,create a new medical business model. We have been collaborating with EBM to develop a UI based on the PACS system. In the field of AI, it can improve international collaboration of AI development through federated learning,can facilitate the establishment of local AI industry, from development to verification by regulatory agency.

天使投資人、策略合作夥伴