Abstract
Artificial intelligence (AI) models are playing an increasingly important role in biomedical research and healthcare services. Artificial intelligence can be used for repetitive tasks such as disease diagnosis, disease screening, disease prevention, and early disease by analyzing and interpreting a big amount of data from different sources and different formats in the hospital. The speech address on exploring the fields of healthcare artificial intelligence, especially in applications and challenges of AI-powered medical image diagnosis, such as in cardiovascular imaging, coronary arteries in CCTA usually have low and non-uniform signal-to-noise ratios (SNR), leading to a great challenge for coronary lumen segmentation. In addition, the coronary arteries were usually annotated incompletely due to wide inter-individual variations, it makes the coronary lumen segmentation and stenosis detection more difficultly. To derive accurate and complete reconstructed coronary arteries, we developed a multi-strategic approach. Another example, evaluation of healing of apical tissues after root canal therapy by algorithms and ML. At the same time, the challenges of clinical validation and landing will be mentioned.