Send "Performing Eyes" and "Smart Brain" to investigate the digestive maze. Behind the scenes, there are clinicians "finger roads", surveying and mapping team drawing, AI engineer "calculation road"

The Yangtze River Daily Da Wuhan client August 17th "High -risk lesions, please observe carefully!" "The remaining 8 parts, please continue to refuel!" When the patient conducts gastroscopy, the "endoscopic elf" automatically identifies the blind area, reminding the doctor's inspection parts that they have not arrived, while reminding the risk lesions in real time to avoid leakage diagnosis of early cancer of the digestive tract.

This is just one of the application scenarios of the "endoscopy" artificial intelligence digestion endoscopy. Wuhan University People's Hospital (Hubei Provincial People's Hospital) Digestive Hospital Professor Yu Honggang's independently developed and leads the auxiliary diagnostic system for digestive endoscopy artificial intelligence. More than 500 medical institutions across the country.

Data show that the incidence of gastrointestinal tumors accounts for 43.5%of the total cancer incidence of Chinese cancer. Endoscopic examination is a necessary means for early detection of digestive tract tumors. Five years ago, Yu Honggang began to study how to use artificial intelligence to touch where the "eye force" and "hand force" could not be reducing the blind spot rate during gastroscopy and improved the diagnosis of early gastric cancer.

Yu Honggang's team found Dr. Hu Shan, the Image Processing Team of the School of Resources and Environmental Sciences of Wuhan University, finally "unveiled the list". There are also more than 30 research teams with artificial intelligence knowledge background and hatched Wuhan Chu Elf Medical Technology Co., Ltd.

Dr. Hu Shan said that due to the clinical application orientation of product research and development, this creation did not stay on "paper and talk about soldiers". The efficiency of the combination of medical workers was much higher than that of engineers.

The team cooperated with the School of Surveying and Mapping of Wuhan University to draw the "topographic map" of the human digestive tract. In accordance with the diagnosis and treatment specifications of digestive endoscopy, the "coordinate point" that must be checked. The team marked the stomach as 26 parts, scanned by artificial intelligence technology one by one to accurately identify to ensure that there is no blind zone; at the same time, imitate learning expert diagnostic logic and build a logical anthropomorphic early gastric cancer identification model.

Wuhan Artificial Intelligence Computing Center also joined. Hu Xiao, deputy director of the technical deputy director of Wuhan Chu Elf Medical Technology Co., Ltd., introduced that in order to achieve the ideal accuracy performance of deep learning models, big data samples need to be provided for repeated training and tuning. Take early gastric cancer identification model as an example. The super -strong AI computing power, more than 100,000 data set training time from 1 week to 1 day. Relying on the self -innovating soft hardware platform of the Wuhan Artificial Intelligence Computing Center, the learning process of the "Elf" can ensure information security.

Can you really compete for a real knife? The team took the "Endoscopy Elf" to participate in the Early Gastric Cancer Diagnosis Competition in the Second Elf Cup. In this competition that lasted 3 hours and "who can accurately draw early gastric cancer", the "endoscopic elves" defeated the 30 digestive endoscopes across the country. The key front -line tool, the image data of white light endoscopy, to determine early gastric cancer, it takes 30 minutes to watch 100 videos and make judgments. "Endoscopy Elf" is witty and win first.

Hu Xiao introduced that after more than 400 versions of the latest products, it was released in July. The average accuracy of monitoring gastroscopy was 95.3%, and the accuracy of early gastric cancer recognition was 90.2%. Room for improvement.

Recently, the Department of Science and Technology of the Ministry of Industry and Information Technology and the Medical Device Registration Department of the State Drug Administration announced the unveiling unit of the innovation task of artificial intelligence medical device for artificial intelligence medical device. , Accelerate the application of new technology and new products, led by Wuhan Chu Elf Medical Technology Co., Ltd., Wuhan University, Wuhan University People's Hospital, Zhejiang University Medical College Hangzhou First People's Hospital, China University of Geosciences (Wuhan) jointly declared " "Gastrointestinal endoscopy imaging auxiliary diagnostic software" project was shortlisted for the unveiling of artificial intelligence medical device innovation tasks.

【Link】

Artificial intelligence medical device innovation task unveil list

Wuhan shortlisted number is the forefront of inland areas

The unveiling unit of artificial intelligence medical device innovation tasks announced by the Ministry of Industry and Information Technology. There are 11 projects on the list in Wuhan, ranking among the top inland areas.

The reporter from the Yangtze River Daily observed that among the shortlisted projects distributed in the city, Beijing, Shanghai, Hangzhou, Chengdu, Shenzhen, and Wuhan.

Wuhan shortlisted units are Wuhan Chu Elf Medical Technology Co., Ltd., Wuhan Lanting Intelligent Medicine Co., Ltd., Solid Wandosheng Medical Technology (Wuhan) Co., Ltd., Anhan Technology (Wuhan) Co., Ltd. Company, Wuhan Zhongqi Biomedical Electronics Co., Ltd., Huazhong University of Science and Technology Tongji Medical College Affiliated Xiehe Hospital, Wuhan University, Wuhan Lianying Zhicong Medical Technology Co., Ltd., Huazhong University of Science and Technology, Wuhan University People's Hospital.

It is worth noting that the Wuhan unveiling project project has both enterprises and colleges and medical institutions. It is mainly distributed in the field of tumor treatment, cardiovascular intervention treatment, intelligent monitoring and other fields. In the future, the state will increase support for unveiling units in the aspects of medical device review, scientific research projects, and application pilots, and accelerate the application of the results of the unveiling list.

Just in the near future, the six national departments issued the "Guiding Opinions on Accelerating Scene Innovation to Promote the High -quality Development of Economic High -quality Economic Development", which proposed "strengthening the role of corporate scene innovation, encouraging the participation of colleges and universities to participate in scene innovation, cultivating and strengthening scene innovation Professional institutions, building a new highlands of artificial intelligence scenes ", etc.; The Ministry of Science and Technology announced the" Notice on Supporting the Construction of a New Generation Artificial Intelligence Demonstration Application Scene "on the 15th. Intelligent diagnosis and treatment, as one of the top ten scenes supporting, were ranked in it. According to the person in charge of the Wuhan Artificial Intelligence Computing Center, the units of the unveiling of the artificial intelligence medical device innovation task of the Ministry of Industry and Credit Ministry of Industry and Credit Ministry of Industry and Credit Ministry also have Coppert. The second is shortened to 3 seconds, which is also applied to the center of the center's computing power support.

It is understood that around artificial intelligence+medical/diagnosis and treatment, Wuhan has 38 three hospitals, 18 national medical centers approved nationwide, and 29 clinical specialists rank among the top ten in the country. At the same time, Wuhan has many universities and enterprises in this field. There are also new infrastructure and new scenarios, and the "intelligent diagnosis and treatment" ecosystem will accelerate.

In the field of intelligent diagnosis and treatment, the national artificial intelligence medical device innovation tasks shortlisted in Wuhan unveiled list reflect the contemplates of Wuhan medical institutions, enterprises and new infrastructure in building new scene goals. Seven years ago, this was just a doctor's idea, and now it has provided more than 500 medical institutions across the country.

(Changjiang Daily reporter Li Jia Correspondent Wu Jing Xuan Chi Xia Yang Jiani)

【Edit: Ding】

For more exciting content, please download the "Da Wuhan" client in the major application markets.

- END -