Core demands
At present, the general responsibility of the radiology department in domestic hospitals is to provide diagnostic services for the images of the whole hospital, of which the chest image is the most conventional and inexpensive examination item, which leads to a large number of examination results in hospitals every day requiring the radiology department to diagnose.
Generally, after the radiologist in the general hospital has issued a request for image examination and diagnosis, the patient needs to come to the radiology department first to complete image acquisition (such as X-ray photography) by the technician operating the equipment, and then submit it to the trained specialist (radiologist) for diagnosis. However, if the number of patients is large on a certain day, doctors need to work with the display every day, which will directly lead to the naked eye fatigue of doctors in the face of high-intensity work, and the phenomenon of misdiagnosis and missed diagnosis will also be difficult to avoid.
If artificial intelligence technology can be applied to auxiliary diagnosis, doctors can make preliminary judgment processing faster with the aid of AI, so as to reduce labor intensity and improve diagnostic accuracy. Therefore, there are the following urgent problems in domestic hospitals:
1. In reality, the radiologist will spend 10-30 minutes waiting time when observing and making a diagnosis with the naked eye, which will pose a high test to the radiologist's work intensity and experience requirements.
2. In addition, there is a shortage of radiologists in secondary hospitals and township hospitals. If automated AI aided diagnostic equipment can be used, it will be possible to greatly improve the diagnostic efficiency between doctors and patients.
Solution
Guangzhou Kaihui is a cooperative customer of Baidu AI. Its solution is to first configure an intelligent diagnostic terminal with Baidu EasyDL customized image recognition technology, then connect it with the hospital image archiving service system, so that the images taken by patients can be acquired in real time, and then go through systematic classification processing (similar to triage) Then it will be directly transferred to the AI diagnosis server (such as lung cancer model), and finally the feedback results will be timely transferred to the doctor workstation after the diagnosis is completed.
In addition, in the case of a lack of radiologists in a secondary hospital, the intelligent diagnostic terminal will be directly connected to the imaging equipment, and will also be sent to the diagnostic server through classification processing, and then the results will be fed back to the doctors.
Highlight 1: Optimized AI aided diagnosis process
Highlight 2: AI server is connected to EasyDL server, which can obtain the image taken by the patient in real time
Highlight 3: AI diagnostic workstation is connected to the server to quickly transmit diagnostic information to the doctor workstation
Highlight 4: Automatic classification of normal/lung cancer/pneumonia, and feedback the results to doctors
