The prosperity of artificial intelligence has become one of the new driving forces to promote social and economic development. It plays an important role in improving social production efficiency and realizing social development and economic transformation. As the core force leading the new generation of industrial transformation, artificial intelligence has shown new applications in the medical field and spawned new formats in the deep integration.
In fact, compared with manufacturing, media, retail, education and other fields, AI is still in the early stage of medical treatment, with relatively low commercialization and low industry penetration, which is closely related to the nursing and conservatism of the medical industry. However, it is undeniable that the combination of artificial intelligence in the medical field has responded to many difficulties of traditional medical treatment, with wide market demand, diverse business trends and broad development space.
New crown disease promotes artificial intelligence from the cloud to play a key role in improving the overall efficiency of anti epidemic. Epidemic has become the touchstone of artificial intelligence in the medical field, showing the strength and value of artificial intelligence in medical treatment. From the perspective of application scenarios, AI medical application is still in its infancy, with image recognition, remote query and health management ranking the first echelon.
Among them, image recognition, as a subdivision field of auxiliary diagnosis, is the most widely used scene of artificial intelligence in the medical field.
The concept of imaging diagnosis and treatment originated from the field of oncology, and then extended to the whole field of medical imaging. Understanding medical images and extracting key information with diagnostic and treatment decision-making value is a very important link in the process of diagnosis and treatment.
In the past, 4-5 doctors were required to participate in the diagnosis of medical image preprocessing. However, based on artificial intelligence image diagnosis, training computer to analyze medical images, only one doctor participates in quality control and confirmation, which has great benefits to improve the efficiency of medical behavior.
Artificial intelligence first explodes and lands in medical images, mainly because the access and processing of image data is relatively easy. Compared with medical records and other data accumulated for more than three or five years, the image data can be obtained in a few seconds with only one shot. An imaging film can reflect most of the patients’ conditions and become the direct basis for doctors to determine the treatment plan.
The huge and relatively standard database of medical images and the continuous progress of intelligent image recognition algorithm provide a solid foundation for the application of artificial intelligence medicine in this field.
From a technical point of view, medical image diagnosis mainly depends on image recognition and deep learning. According to the clinical diagnosis path, firstly, the image recognition technology is applied to the perception link to analyze and process the unstructured image data and extract useful information.
Secondly, using deep learning technology, a large number of clinical image data and diagnosis experience are input into the artificial intelligence model to make neural network carry out deep learning training. Finally, based on the algorithm model of continuous verification and grinding, the intelligent reasoning of image diagnosis is carried out. Output personalized diagnosis and treatment judgment results.
The combination of artificial intelligence based on image recognition and deep learning with medical image can solve at least three needs. 1、 Focus recognition and annotation, that is, medical image segmentation, feature extraction, quantitative analysis, comparative analysis, etc. To meet this demand, the automatic recognition and marking system of X-ray, CT, MRI and other medical images can greatly improve the diagnostic efficiency of imaging doctors. At present, Al medical imaging system can quickly process more than 100000 images in a few seconds, improve the diagnostic accuracy, especially reduce the false negative probability of diagnostic results.
2、 Automatic target delineation and adaptive radiotherapy. Automatic target mapping and adaptive radiotherapy products can help radiologists automatically sketch 200 to 450 CT images, greatly reducing to 30 minutes. During the 15 ~ 20 times of irradiation, the location of lesions can be continuously identified to achieve adaptive radiotherapy, which can effectively reduce the damage of radiation to patients’ health tissue.
3、 Three dimensional image reconstruction. The registration algorithm based on gray statistics and the registration algorithm based on feature points can solve the problem of fault image registration, save the registration time, and play a role in the location, scope, benign and malignant lesions recognition, surgical scheme design and so on.
From the perspective of landing direction, the current layout of Al medical imaging products in China mainly focuses on the chest, head, pelvis, limbs and joints, and mainly focuses on the leading cities of cancer and chronic disease screening.
In the early development and application of artificial intelligence medical imaging, pulmonary nodule and fundus screening is a hot field. As technology matures and iterates over the past two years, major al medical imaging companies are expanding their business scope, and bone age testing around breast cancer, stroke and bone joints has become a key area for market participants. Aluminum medical imaging novel coronavirus pneumonia is a key factor in improving the diagnostic efficiency and quality of diagnosis.
Dual entry of policy capital
If the relative accessibility and processing of image data is the main reason for the first outbreak and landing of AI in medical images, then the support of national policies and a large number of capital access endow AI with the power of continuous updating in the application of medical images.
From the perspective of policy addition, from 2013 to 2017, various government departments issued a number of policies to continuously increase the support for domestic medical imaging equipment, third-party independent medical imaging diagnosis center, telemedicine and other fields.
At the end of 2016, the State Council issued the “13th five year plan” national strategic emerging industry development plan, which mentioned medical imaging for many times, and pointed out that it was necessary to “research and develop high-quality medical imaging equipment”, “support enterprises, medical institutions, research institutions and other joint construction of third-party imaging center”. In January 2017, the national development and Reform Commission included medical imaging equipment and services.