In recent years, artificial intelligence has continued to land in the home, security, medical, sports and other fields. Since 2020, due to the catalysis of new coronary pneumonia, people have become more aware The importance of health. With the entry of many companies, the technological wave of “AI+sports+medical” is gradually changing people’s daily habits. Among them, as important hardware devices for digital health, wearable devices represented by smart watches and smart bracelets have huge On March 16, IDC announced the sales of smart wearable devices in 2020. In 2020, 444.7 million wearable devices were sold globally, a year-on-year increase of 28.4%. Among them, the shipments of smart watches and smart bracelets accounted for 24.1% and 11.5% respectively, and the two together reached 158 million units.

Pictures are provided by the author, and the copyright belongs to the author

AI sports data collection

Common exercise data includes exercise distance, moving speed, number of steps, stride length, time, route, calories burned, etc. The intelligent Wearable devices can basically achieve real-time monitoring. So how does the smart bracelet realize these functions? Take the step counter function as an example: the human body wears a sports bracelet and measures the walking direction and acceleration force through an acceleration sensor. It can determine whether the device is in a horizontal or vertical position to determine whether the device is moving. The data collected by the sensor is relatively rough, and the MCU The processing error will be very large, so data processing is required, and then the signal is processed through a certain algorithm to realize the basic step-counting function. However, factors such as wearing position, movement, and walking speed differences may affect the accuracy of the bracelet monitoring. In order to make the data of the sports bracelet more accurate, manufacturers usually need to manually collect a large amount of exercise data, such as a handheld pedometer as a control group. , Calibrate with the bracelet data to further improve the equipment and improve the algorithm.

AI medical data collection

Not satisfied with simple exercise data, more and more smart bracelet watches have added heart rate, blood pressure, and blood oxygen monitoring functions. Huawei’s latest HUAWEI WATCH GT 2 Pro, provides intelligent heart rate monitoring, 7*24 hours pressure monitoring, and all-weather blood oxygen saturation monitoring functions. Let's take blood oxygen monitoring as an example: the measurement of blood oxygen saturation mainly relies on the SpO2 sensor built into the smart watch. The absorption ratio of hemoglobin and oxyhemoglobin to infrared light and red light in the blood is different. Two LEDs of infrared light and red light are used to irradiate the finger at the same time, and the absorption spectrum of the reflected light can be measured. The blood oxygen content can be measured by using these two frequencies. With the light, the optical sensor of the smart bracelet can detect the pulse data optically, and calculate the user's blood oxygen saturation based on the data and according to the big data model and algorithm.

Similarly, how to determine that the heart rate, blood pressure, and blood oxygen data recorded by the smart watch are accurate enough?

In the early stage, it still needs to be calibrated with the medical data measured by the hospital’s professional equipment. It is relatively inconvenient for companies to obtain medical data. They need to connect with the hospital's medical record room. It takes time and effort to retrieve the medical record number, and cannot guarantee the completeness of patient information. What is more troublesome is that it involves ethical review.

Data service companies can help companies collect sports and medical data efficiently and quickly, freeing companies from the work of collecting data and focusing on R&D. Jinglianwen Technology has done a lot of data storage and resource storage in the medical AI data collection section, and has collection experience and accumulation of exercise data and medical data such as blood pressure, heart rate, and blood oxygen.

There is no doubt that Jinglianwen Technology has its collection capabilities. The company was established in 2012 and has implemented a large number of collection projects across the country. It has a wealth of hospitals, equipment, and collections.者 Resources. It has maintained long-term and stable cooperation with many hospitals across the country, established a joint collection working group, and has rich experience in ethical review. It maintains long-term friendly cooperation with many sports colleges, primary and secondary schools, gyms, stadiums, and swimming pools, and has accumulated resources for professional athletes and sports enthusiasts in sports colleges, including indoor fitness, ball games, martial arts, dance, track and field, and swimming. Various sports, including multiple age groups and people of color, have collaborated with more than 20,000 model resources to participate in data collection.

In the data headquarters—Hangzhou Bay Information Port has an office area of ​​more than 2,000 square meters, and can open separate collection laboratories at any time according to customer needs, in addition to the whole country There are multiple secondary collection points in the area, and on-site collection services can also be provided according to customer needs.

As a well-known data service provider in the industry, Jinglianwen Technology has a stable collection and management team, which can form a stable supervision team of multiple people at any time according to the needs of customers. Most project managers have participated in the implementation of medical data collection projects, and have extensive experience in coordinating resources, managing projects, controlling progress, and communicating with Party A. At the same time, the project assistant assists in organizing the collection of persons to sign the authorization letter and instructing the collected persons. Carry out collection, complete data sorting and other tasks to ensure the efficient landing of large-scale collection projects.