FU：AI has great potentials in improving medical care, which still requires continuous innovation and development. Although AI may not be the solution to everything, in certain medical diagnosis AI can do much better jobs than humans. With deep learning, AI algorithms can distinguish small features and correlate different types of images or models, which are hard to be done by naked eyes. I would like to share two examples of what AI can achieve. 人工智能在医疗领域的应用仍需不断的创新和发展，同时潜力巨大。尽管目前人工智能还无法解决全部问题，但是在人工智能医学诊断方面已经表现出比人类更出色的能力。运用深度学习技术，人工智能算法对于医学影像的微小的、高阶的变化比人类敏感，在这里我想和大家分享两个相关案例。
The first example is that AI is good at correlating images between different tests, such as MRI and CT scans. In the diagnosis of certain cases of stroke, MRI images have a high single-to-noise ratio but hard to acquire. On the other hand, CT images have a low single-to-noise ratio but easy to acquire. By AI deep learning, it is possible to use CT images to generate virtual MRI images which can provide an accurate and timely diagnosis. 第一个例子是人工智能善于将不同测试之间的图像进行关联，例如MRI和CT扫描。比如说对于某些缺血性脑卒中MR影像为金标准是因为其拥有高S/N比，但却很难获得。相反的，CT的影像虽然S/N比低，却容易获得。利用深度学习我们可以基于CT的影像模拟出MR的影像，这是人类无法做到的。
The second example is about the power of AI in data integration. In the diagnosis of coronary artery diseases, we need to obtain measurements, such as blood flow rate, pressure, wall shear force, and calculate a parameter as known as Fractional Flow Reserved (or FFR) based on the CT images. Those comprehensive measurements require integration of image recognition, segmentation, and computational fluid dynamics. For this kind of multidisciplinary task, AI is much more powerful than mankind. 第二个例子是关于多模态的综合分析。比如我们无创冠心病筛查项目中，人工智能从CT影像中能自动分割提取冠状动脉、自动网格划分、计算血流动力学，从而模拟得到冠脉上任何一点流速、压力、wall share force与最后计算出Fractional Flow Reserved, FFR。这种结合影像分割、流体动力学的多模态定量分析是人类无法做到的。
Besides the strength of AI, one important task is how to leverage the power of AI and doctors to improve the workflow of medical diagnosis.
FU：The first is about the timely diagnosis. China has over a million cases of stroke each year. For stroke patients, time is brain and AI can greatly improve medical care in such cases of emergency. Although the gold standard for some stroke diagnosis is using MRI, it often takes days of waiting for available machines. CT scanners, however, are standard and readily accessible settings in most of stroke centers. Therefore, our CT-based virtual MRI technology will ease the time pressure by providing a precise and fast diagnosis that is shortened from days to an hour. 首先，对于时效性要求很高的疾病。如脑卒中 time is brain（最好能小于6HR），AI能在最短的时间内辅助医生(尤其是缺乏经验的)确诊脑卒中类型，进而采取正确的治疗方案，对于中国每年新发患者超过百万的脑卒中病患能及时提供正确的诊疗建议。
The second revolution AI can bring us is to facilitate universal and affordable health care. One good application of AI-assisted diagnosis is to replace invasive tests with non-invasive ones. In the traditional diagnosis of coronary heart diseases, we need to insert a pressure sensor into a coronary artery to obtain FFR value. It is invasive and expensive- a sensor costs about ten-thousand RMB. So, we believe the CT-based FFR technology will provide non-invasive and low-cost diagnosis and help more patients. 人工智能带来的第二个革命是为我们带来了可负担的医疗服务。人工智能辅助诊断的一个优势在于无创。在传统FFR测量是侵入式的压力导丝，不仅费用昂贵且对患者的心理压力也大。如今人工智能的FFR CT不仅价格便宜，因其非侵入式的优点老百姓的接受度相对也会比较高，对于冠心病筛查的广谱性推广有很大的帮助。
Aitrox Tech has cooperated with tens of hospitals for the R&D and clinical tests on these advanced AI technologies. We hope soon to launch these tools that allow more patients to receive better health care, especially in regions lacking medical resources. And these are some of the great benefits AI can bring to the world. It’s my pressure to be here and thanks for the opportunity to share our vision with the listeners. 以上两个例子，我们杏脉科技正在与几十家医院合作R&D与临床实验，相信很快会应用与帮助到老百姓，尤其在医疗资源匮乏地区。这些都是人工智能给世界带来的一些益处。再次感谢有机会与大家分享科技赋能医疗的愿景。