"How does the internal medicine go?"
"Which medical department do you want to find? Or do you want to go to the internal medicine inpatient building?"
"I am looking for a digestive medicine."
"Please go to the hall to take the escalator or elevator to the north side of the fifth floor of Zone B. For the detailed location of the department on the fifth floor, please see the map below."
This conversation took place in the lobby of the General Hospital of the People's Liberation Army not long ago. Mr. Zhao, a Beijing citizen who came to see a doctor for abdominal pain, was inquiring about the location of the medical department to the intelligent medical consultation robot “Xiao Medicineâ€.
"I didn't expect this robot's reply to be quite accurate." Mr. Zhao followed the path of "Xiao Medicine" and quickly found the visiting department.
In March 2017, six intelligent guided robots developed by Keda Xunfei were launched at the People's Liberation Army General Hospital, four of which were placed in the outpatient hall. Every morning, these lovable robots are surrounded by people who come to the consultation. Some people are out of curiosity. Some people have something to do. For the questions people ask, the robot will answer them patiently and meet people who are joking with it. It can also make people boring.
“Now they are busy. Every robot has to answer more than 900 questions on average every day. Patients want to find departments, find wards, and consult life services around the hospital. Robots can provide guidance.†Director of Outpatient Department of PLA General Hospital National hi introduces.
"Today's AI is like an Internet, and it has become an infrastructure." Chang Jia, the head of Tencent's "Internet + Medical", said that in recent years, many companies involved in AI at home and abroad have accelerated their deployment to the medical field. From concept to reality, in addition to intelligent consultation, there are also smart doctors who can help doctors see films and do screening.
In the endoscopic operation room of the Affiliated Tumor Hospital of Sun Yat-sen University, the patient lies on the operating table, and the doctor operates the endoscope to extend into the patient's esophagus to collect photos. These photos were transmitted to the computer in real time, and then connected to the AI ​​system. Ten seconds after the end of the photo, the computer showed a line of text: "The suspected cancer risk is 55%." Based on this result, the doctor evaluates whether further pathological biopsy is needed. To avoid missed diagnosis of cancer.
This line of text is a cancer risk warning from Tencent's medical imaging AI system "Tencent Shadow". "It is very convenient to use, just like the doctor's third eye." Xu Guoliang, director of the Department of Gastroenterology, Affiliated Tumor Hospital, Sun Yat-sen University, said: "There was a doctor who checked the endoscopy of the patient and looked at the patient. Later, AI reminded the patient that the esophagus was abnormal. The doctor asked someone to come back for a review. It was really early cancer."
Have the ability to learn and are solving the medical industry problems
It takes more than ten years to train an attending physician, and it takes at least one year to train a skilled doctor. What is the AI's ability to stand in such an industry?
“Intelligent guidance robots are not the same as electronic inquiry machines that have been distributed in public places a few years ago. They have the ability to learn.†The country said that the electronic inquiry machine may have been “cracked†when encountering problems that cannot be answered, but the intelligent robot passed Accumulate and update data, and be able to continuously answer new questions raised by people.
This ability to bring together and apply experience is solving some of the challenges in the medical industry.
The first is to efficiently fill the huge and complex information service needs in hospitals.
The total number of outpatients in the PLA General Hospital is about 18,000, of which 70% are foreign patients. Many people come here to see a doctor. Is it a doctor or a registration first? Where are the doctors' departments? Which building is hospitalization and surgery? “Every time I meet a 'white coat', the patient asks, but the guide is not the doctor’s main business. They don’t know the answer every time. Therefore, the needs of the masses for information services have not been well met for a long time. "The country likes to say.
In the past, the hospital's solution was to establish a medical team. In 2014, the PLA General Hospital recruited 14 full-time medical doctors in the outpatient building and hired 17 hours of work, but still could not meet the huge amount of consultation. “After the on-line intelligent robots are on the line, it is a good complement to the problem of information demand and supply imbalance.†The country said, “They react quickly and tirelessly, and even the accented Mandarins like 'Chuanpu' and 'Guangpu’ understand. This is practical and can better help patients in the field."
The second is to reduce misdiagnosis and missed diagnosis during doctor's examination.
Yan Weiwu, an associate professor in the Department of Automation at Shanghai Jiao Tong University, said that in the medical field, AI has been applied to information consultation, medical imaging, electronic medical records, etc., among which medical imaging has the fastest progress.
"China's esophageal cancer is high, but the diagnosis of early cancer lags behind that of developed countries." Xu Guoliang analyzed that there are two main reasons for the low rate of early cancer detection in China. On the one hand, doctors generally have a large workload, and it is difficult to ensure the fine verification of patients; Early cancer is difficult to detect, such as early esophageal cancer. From endoscopy, the symptoms often show local congestion, mucous membrane roughness or subtle erosion. It is very similar to general inflammation, and some inexperienced doctors can easily ignore it.
"When we know that AI can learn to look at endoscopic photos, we are very welcome, because AI can make up for the doctor's subjective negligence." Xu Guoliang said that at the end of 2016, Sun Yat-sen University Cancer Hospital desensitized more than 100,000 pieces. The endoscopic photo of the esophagus is provided to Tencent for scientific research. The AI ​​learns a large number of photo data through deep learning to form a set of examination standards, which can significantly improve the rate of early cancer detection.
In the People's Hospital of Zhejiang Province, Tencent Yingying has been used for screening and discovery of esophageal cancer, diabetic retinopathy, and pulmonary nodules (used to diagnose early stage lung cancer). In the experimental environment, the AI ​​system has an accuracy rate of more than 90% for the first two diseases and 85% for the pulmonary nodules.
Huang Dongsheng, dean of the People's Hospital of Zhejiang Province, believes that using AI to analyze image data such as chest radiographs, CT, and pathological sections has improved the efficiency of reading films and freed doctors from hard work. "This helps doctors to devote themselves to the rare diseases and incurable diseases that have not been learned in the past. These are the machines that cannot be learned at present."
Increased social recognition and enhanced technical applicability
Chang Jia told reporters that Tencent Yingying has been launched since August 2017. It has established a “Artificial Intelligence Medical Joint Laboratory†with more than 30 top three hospitals nationwide. “Today, AI’s penetration into the industry is increasing and social recognition is improving. Technical applicability is increasing."
But at this stage, the prospects of medical AI are not all the way.
“Technically, the compatibility of medical imaging equipment with AI systems is a problem. Different manufacturers have different camera equipment, data formats and picture quality, which interferes with the accuracy of machine learning. In addition, the AI ​​database of hospitals needs to be fine. Fine-tuning, the accuracy of AI diagnosis may be affected by the region, the same disease will show small data differences in different regions, the technician is solving." Chang Jia said.
"In terms of user habits, the people still stay in the touch-based mode for human-computer interaction, and are not accustomed to talking directly with the robot. Some people do not trust the machine and only choose to resort to people. This means that the society must fully accept new technology. It still takes time." The country said.
On the macro level, Wei Weiwu pointed out that the development of medical AI faces the following obstacles: First, industry recognition, doctors have a wait-and-see attitude towards AI, this new technology has not yet fully gained the trust of the medical community; second, data resources, policy has not yet It is completely open, and the issues of privacy and ownership of residents' health data need to be clarified. Third, legal ethics, AI is neither traditional medical equipment nor human beings. How is it technically certified? How to determine the responsibility for medical malpractice? This is worth discussing. After all, the accuracy of AI assist has not reached 100%.
"The most important thing is the interpretation of medical AI." Wei Weiwu said, "Medical is a serious science. Diagnosing illness has a set of traditional logic, but AI is challenging this logic. Even though it can reach or even surpass doctors' diagnosis. The accuracy, but it is based on the diagnostic principle of big data and doctors use medical knowledge to make a diagnosis is not the same, the interpretability of this new technology different from the previous medical law, affects the acceptance of AI in the medical community. â€
Although there is still a long way to go, many industry insiders are optimistic about the development of medical AI.
"Comprehensively, China's medical AI level is second only to the United States, and its development is very fast. It has achieved outstanding results in many AI competitions. We are internationally leading." Wei Weiwu said that the Ministry of Industry and Information Technology had previously issued "Promoting a new generation of artificial intelligence industry." In the development of the three-year action plan (2018-2020), among the eight types of intelligent products that are mainly cultivated, there is a medical image-assisted diagnosis system. "The launch of the action plan, given the timetable, gives the development path, and specific requirements have been put forward. I think that industrialization will soon be realized in this field."
"Some people are worried about the safety of AI. From the perspective of hospital administrators, I feel that in the face of a new technology, we can't resist it from the beginning, but we must first understand it. If the verification is good, then I should take the initiative to embrace.†Huang Dongsheng said, “With the continuous cooperation of doctors and technicians, I believe that medical AI will be more and more intelligent, and will provide people with a higher level of service in the future.â€
We make OBD connector with terminal by ourselves, soldering type and crimping type are both available. Also 12V and 24V type. OBD1, OB2, J1939, J1708, J1962, etc. Also molded by different type, straight type or right-angle type. The OBD connector cables used for Audi, Honda, Toyota, BWM, etc. We have wide range of materials source , also we can support customers to make a customized one to replace the original ones.
Vehicle Diagnostic Cables,Diagnostic OBD Cable,OBD2 Splitter Y Cables,OBD2 Diagnostic Adapters,OBD Heavy Vehicle Cables
ETOP WIREHARNESS LIMITED , https://www.oemmoldedcables.com