The film "I am not a drug god" is not only a hot topic, but also a hot discussion about the over-pricing of patented drugs. The application of AI technology in the medical field is once again discussed. In a demonstration of lung cancer CT image artificial intelligence system, more than 10 images were detected, and a small red circle appeared on the image after uploading the system through the cloud. The circled range is the lesion automatically recognized by the system, and the accuracy rate is 93. .9%. In addition to recognition speeds higher than young pathologists, accuracy is also high. And this is just the tip of the iceberg of AI medical imaging applications. According to Huihui Huiying, only through the intelligent initial diagnosis system platform, breast cancer imaging diagnosis can save 60-70% of the time, chest radiograph diagnosis can save nearly 50% of the time; in the accuracy rate, self-iteration through various hospital data And the model is precisely tuned, and it has reached more than 95%. In the past two years, with the development of technologies such as machine learning and deep learning, AI systems based on pattern recognition have surpassed human performance in both image and speech, especially in the fields of security, finance and medical. Among them, in the past year, China AI Medical announced more than 30 financing incidents, with a total financing of 1.8 billion yuan. The application in the medical field is mainly AI imaging. Due to the maturity of technology and capital catalysis, the AI ​​imaging industry has also begun to explode. In an interview with reporters, Huiying Huaying CEO Chai Xiangfei said that there must be two conditions for the outbreak of any industry. First, market demand and technical conditions, both of which can form an industry. "Reviewing AI images for three years is quite interesting." Since 2015, anti-AI represented by Bill Gates and Musk, the main camp is clamoring for AI technology to commercial applications, and then to the first year of industrialization this year, AI technology has gradually matured and national policies have been blessed. "This is a very good era and entrepreneurial environment." Chai Xiangfei told reporters that it is in this context that he decided to return to China to start a business. Chai Xiangfei is a Ph.D. in medical imaging at the University of Amsterdam and a postdoctoral fellow at Stanford University. He has over 10 years of interdisciplinary research experience in medical imaging artificial intelligence, image processing and data analysis. In 2015, after returning to China, he founded Huihui Huiying. At present, the company has formed a mature medical association hospital solution, which has created digital, mobile and intelligent medical images, and completed decision support from screening, diagnosis and treatment. closed loop. “The original intention of the founding of Huihui Huiying was to develop an AI imaging product that the hospital can really use, to solve the shortage of imaging doctors and a large number of missed diagnosis and misdiagnosis problems, especially in many second- and third-tier cities and towns in China, the people can not get very good. The number of imaging doctors can't keep up with market demand. The high efficiency and high accuracy of artificial intelligence systems give them great hope and ability to completely solve this problem and free the hospital from the heavy dependence on imaging doctors. †Everything seems to lay the foundation for the next event. In 2015, shortly after the company was founded, Chai Xiangfei suffered awkwardness. At that time, the degree of informatization of Chinese hospitals was seriously lagging behind. Even in the top three hospitals in the first-tier cities, the proportion of “upper clouds†in IT systems was very low, and the degree of cloudization in American hospitals at that time was already high. Chai Xiangfei decided to start from the image cloud platform, mainly to solve the problem of the bottom of the hospital, so as to open up the data connection between the hospital and the hospitals at all levels. This also plays an important role in the grading diagnosis and treatment advocated by the state. It is a medical association and region. The important infrastructure of the imaging center. Take the North Hospital Third Hospital as an example. After using the intelligent image cloud platform, the IT system realizes “upper cloudâ€. In addition to accumulating a lot of engineering experience for Huihui Huiying, this cooperation also laid a good data foundation for the later AI imaging system. In 2016, Huiyi Huiying received a 10 million-level A round of financing from Lanchi Ventures, which is also the first medical image financing in the industry. Some people commented that this financing has stimulated capital's enthusiasm for investment in AI medical care. Because in the near future, companies such as Yasen Technology, Lianxin Medical and Imagination Technology have successively received more than 10 million investment, no matter the amount of investment or time interval, it is not difficult to see the "AI+ medical imaging" field. In 2017, this is a year of icing on the cake. This year, Chai Xiangfei's mentor Xing Lei officially joined as the company's chief scientist. Xing Lei is a tenured professor at Stanford University and the head of the Department of Medical Physics at Stanford Medical College. He is also a professor of electronic engineering, molecular imaging and bioinformatics, and Professor Bio-X. He has been teaching medical imaging, medical physics and medical information for more than 20 years. It is also an expert of the national “Thousand Talents Programâ€. Under his guidance, Huihui Huiying systematically and systematically introduced the academically popular image data analysis method "radiology omics" and further introduced it to the market. Radiology can be applied to many projects such as disease detection, lesion segmentation, diagnosis, treatment selection (personalized medicine), efficacy evaluation and clinical prediction. At present, Huiping Huiying's radiology group Yunping has been stationed in more than 300 hospitals in China, and is used by more than 1,000 radiologists. It can provide one-click analysis of lesion delineation, eigenvalue calculation analysis, machine learning and case prediction. Detailed quantitative reports were issued for different diseases. "Of course, if radiology + AI stops at scientific research, the volume is too small and the value is not big. We hope to explore a set of methods to allow doctors to participate in medical innovation and translate scientific research into clinical validation and application." Chai Xiangfei said. For example, Huihui Huiying and Director of Department of Vascular Surgery, Beijing 301 Hospital, jointly developed the aortic artificial intelligence research cloud platform AORTIST2.0, which solved the accurate measurement, prognosis prediction and remote follow-up in type B aortic dissection. The core problem is the world's first automatic segmentation method for B-type aortic dissection artificial intelligence, which was officially released in April this year. At present, Huiyi Huiying products cover the whole chain AI imaging services, including smart image cloud platform, digital intelligent film, tumor radiotherapy cloud platform, big data intelligent analysis cloud platform, image intelligent screening system and artificial intelligence diagnosis cloud platform. Core services, from medical imaging digitization to mobilization to intelligence, complete the closed loop from screening, diagnosis to treatment decisions. “The market cake is bigger, and there is a chance to get more.†In general, the development of a new field, at first, a number of companies each developed the market, after a period of "barbaric growth" before "the narrow road." In the field of AI medical imaging, due to capital boosting, a relatively fierce competitive situation has been formed. In the face of competition, Chai Xiangfei said frankly: "We uphold a relatively open attitude. On the one hand, we have strong technical and product strength, and we are not afraid of competition. On the other hand, there are many competitors, which actually helps to open the market quickly." For hospitals, AI imaging products are new and generally cautious. The visit of many AI imaging companies will help the education market, dispel hospital concerns, increase the market penetration of AI imaging products, and expand the market. “The cake is bigger, and as part of it, there is a chance to get more.†According to the reporter, the commercialization of AI medical imaging companies has always been a common problem in the industry. In the interview, Chai Xiangfei said that the most important thing for commercialization is to clearly define the medical application scenarios of the products, integrate the AI ​​imaging products into the hospital's original information systems and medical processes, reduce the cost of hospital renovation, and create real benefits for the hospital. Only by creating value for the hospital can the hospital accept (saving money). In addition, if AI Medical wants to get through the hospital procurement process, it must also obtain CFDA certification. In August 2017, the State Food and Drug Administration (CFDA) released a new edition of the Catalogue of Medical Devices, adding a category corresponding to AI-assisted diagnosis. According to the regulations, if the diagnostic software provides diagnostic advice through the algorithm, only the auxiliary diagnostic function, and does not directly give the diagnosis conclusion, the second type of medical device is declared; if the lesion is automatically identified and a clear diagnosis prompt is provided, then according to the third Medical device management. However, the second type of equipment has a clinical trial exemption list, and the diagnostic software declaration can be exempted. The CFDA has not yet made specific specifications. The third category of medical devices requires clinical trials. Now, Huiyi Huiying has obtained a second-class medical device registration certificate. There are four types of three types of medical devices at the same time. The approval process has already started the clinical verification phase. Regarding the issue of fees, Chai Xiangfei said that at present, the hospital mainly charges SaaS for service fees, and has already landed more than 700 hospitals. The business revenue is in good condition and is in a stage of rapid growth. Other charging models are also actively explored, such as insurance institutions or C-side charging models. It is worth mentioning that Huiyi Huiying is also actively expanding overseas markets, taking countries and regions along the “Belt and Road†as important markets and providing different products according to local conditions. At the same time, Huiyi Huiying also cooperated with domestic medical device manufacturers to bring soft and hard integrated medical services to hospitals in the “Belt and Road†region. At present, Huiyi Huiying products cover Japan, France, Kazakhstan, the United States, India, Israel and other countries. Among them, Huiyi Huiying signed a contract with Kazakhstan's largest private hospital chain group, and cooperated with Japan's largest cloud pacs company to determine the radiology platform, and cooperated with France's largest oncology company and American AI medical company. Chai Xiangfei also admitted that the overseas market is relatively mature and the competition is more intense, which puts higher demands on the company's technology and market expansion capabilities. "Enterprises must prepare for long-term battles if they want to do AI images." From the current practice of global AI medical companies, the specific applications of smart medical care include: drug development, medical imaging and diagnosis, intelligent diagnosis and treatment, intelligent health management, drug mining, virtual assistants and so on. In the field of AI medical imaging, most enterprises' application scenarios mainly focus on common diseases such as esophageal cancer, lung cancer and breast cancer. In this regard, Chai Xiangfei said that AI medical imaging companies will mainly consider two things in the choice of disease types: On the one hand, the medical value of the disease. Enterprises generally choose a disease with a large number of patients and an improved effect and efficiency after applying the AI ​​imaging system. For example, most AI medical startups have chosen lung nodules at the beginning, because lung cancer ranks first in China in terms of morbidity and mortality, and screening for lung nodules helps prevent lung cancer, so this scenario has great medical care. value. Huiyi Huiying also selected lung nodules at the beginning. At present, the screening accuracy rate of the disease is up to 95%, and the detection rate of 3mm lung nodules is 85%. On the other hand, the medical data availability of a disease type. Through the reporter's understanding, at this stage, the shortage of medical data is also one of the bottlenecks restricting the commercialization of AI medical images. Still taking pulmonary nodules as an example, most companies start with lung nodules because the lung nodule data is relatively easy to find through open networks, such as the LIDC database for chest medical image files (such as CT, X-ray) and Corresponding diagnostic results are labeled. In fact, the scarcity of medical data will limit the speed of enterprise model development. Only with the advantages of multiple disease data, AI imaging companies can realize rapid multi-path model development, which is why many AI imaging companies choose to cooperate with large-scale top three hospitals. The amount is large enough and the data quality is guaranteed. Compared with the second half of 2017, since 2018, we have rarely heard about AI medical imaging companies. It seems that these companies have some "convergence". In the interview, Chai Xiangfei said to reporters at the end of the interview. "Any industry has its own development rules. It is not necessarily a good thing to give too much attention in the short term. After a short period of 'explosive', AI images may experience a period of time. The AI ​​image is an industry that has only begun to erupt in recent years, but it has experienced long-term technology accumulation. If companies want to do AI images, they must prepare for long-term battles." Cooking wine is a type of wine that is specifically made for cooking and not for drinking. It has a higher salt content and lower quality than regular wine, which makes it unsuitable for drinking but perfect for adding flavor to dishes. Bucket cooking wine is simply a larger quantity of cooking wine that comes in a bucket or large container. It is often used in commercial kitchens and restaurants where large quantities of wine are needed for cooking. Bucket Huadiao Alcohol,Huadiao Liquor for Cooking,Huadiao Wine,Bucket Packing Cooking Alcohol Zhejiang Shengta Shaoxing Wine Co., Ltd. , https://www.shaoxingyellowricewine.com