June 16, 2020, HMI (NYSE:HMI) First AI Innovation Conference at University of Science and Technology of China Hefei, Anhui Province, where it was held successfully, the tech company in the Tech University department is once again trying to lead the way.
In the name of AI, Huami is constantly climbing to new heights. In just one hour’s press conference, five AI engines and one optical sensor were released one after another. Of course, the most eye-catching is “Huangshan-2”, a new generation of AI chips in the wearable field.
Today’s Huami has achieved leapfrog upgrades in the three cores of algorithms, sensors and chips. This time, it once again uses technology to connect itself with life and health.
So, what is a truly wearable product that can be used for health management? On top of the conference, Wong Wong described his ideal, wearable health management in three short sentences.
“Exercise should be anytime, anywhere.”
“Records should be readily available.”
“Accumulation should be anytime, anywhere.”
Behind this ideal, Huami has set new requirements for wearable device data recording. According to Huang Wang, “Smartwatch makers are always repeating one thing to users, which is to tell them that their devices can support the Dozens, even hundreds of sports modes, as if the more sports modes they support, the smarter their products are. But if you think carefully about the user experience, this kind of rhetoric seems to be less heartfelt, precisely, the manufacturer’s concern about the direction of the problems that arise. ”
“Real intelligence should be about making the user forget about the ‘settings’ thing, and allowing the device to be really smart about accessing the user’s movement patterns with the information, and give intuitive data feedback – users do not need complex and difficult to understand, non-referential exercise information.”
Clearly, slogans and blueprints can hardly truly connect wearables to the aforementioned health management concepts, and on top of the inaugural AI Innovation Conference, Huami unveiled five AI-based health engines in one go to bring its product design philosophy to life.
RealBeats™ Bio-Data Engine: Analyzes PPG optical heart rate data and ECG electrocardiogram data to enable automatic screening for irregular heartbeat.
Hami Technology has completed a clinical medical study with the First Hospital of Peking University on the smart bracelet to monitor atrial fibrillation. RealBeats™’s smart bracelet PPG and ECG functions determined AF with an accuracy of 93.27% and 94.76%, respectively. This result is consistent with the results of manual interpretation by medical professionals. Thanks to this engine, 9,100 suspected cases of AF have been monitored so far by HamiTech.
This new second-generation heart data AI bio-engine, RealBeats™ 2, may be the world’s most advanced heart rate AI engine. It effectively eliminates noise interference to the heart rate signal during exercise, with night and daytime effective AF monitoring times that are comparable to those of the previous generation. 1.87 times and 6.64 times. In addition, through the establishment of a big data model of heart health, we also successfully achieved the AI-automation of refractory tachycardia and supraventricular frequent premature beats. Screening.
OxygenBeats™ oximetry data AI bioengine: the algorithm pre-processes the oximetry signal based on the health big data model to eliminate signal noise and improve the measurement accuracy by up to 50%; and by using multiple groups of oximetry detection values for calibration, it solves the errors caused by user wearing errors, thus improving the accuracy of the collected information.
Huang Wang introduced, Huami used the oxygen drop experiment to verify the accuracy of the algorithm, the results show that the accuracy of the detection rate can reach 100%, the market! The average error of the test results with professional oximeter is only 1.67%, which is more accurate than most of the similar products on the market. Oxygen detection algorithm for wrist wearable devices.
He revealed that, based on previous collaborations, Huami Technology will also work with a team of academicians from Nanshan Zhong to leverage the OxygenBeats™ High-precision oximetry capability for rehabilitation follow-up of patients with new coronary pneumonia. The smartwatch, which is expected to feature the OxygenBeats™ blood oxygen data AI bioengine, was recently launched this year ( (2020) Q3 release.
SomnusCare™ Sleep Data AI Bio-Engine: the algorithm is based on Huami’s massive health data and enables the Accurate identification of sleep status and multi-dimensional data to help users understand sleep status and quality. It detects sleep data with an accuracy of over 80% and detects duration of over 25 minutes with nearly 100% accuracy. Napping data.
At the same time, for the “invisible killer” sleep apnea syndrome, which is a serious health hazard, Hami Technology combined with OxygenBeats™ blood oxygen engine, can determine the sleep state and blood oxygen saturation detection from the two dimensions of in-depth analysis, to achieve intelligent identification of sleep apnea syndrome, and promptly remind users to take the necessary medical measures.
Previously, during this new crown epidemic, using the biological data engine, and combined with external factors such as weather, seasons, historical cycles, etc., Huami Technology has already established a set of predictive models for epidemic incidence trends. With the full upgrade of the bio-data engine, future epidemic prediction and even early warning capabilities will also become the new health infrastructure platforms’ Part One.
Human activity pattern AI recognition engine ExerSense™: as mentioned above, almost every smart now on the market The watches, all require users to manually set an exercise mode before they can start recording their workouts. The Huami technology analyzes the exercise data and heart rate data from different exercise modes
The engine can detect data from the motion sensor and heart rate sensor on the Huami wearable device to match the exercise model in real time and finally The ability to intelligently determine the user’s current exercise pattern. Currently, ExerSense™ is capable of automatically recognizing 19 movement modes, including walking, running, cycling, and Swimming, etc., covering 95% of the user’s daily sports scenes; the user does not need to carry out tedious manual operation, to achieve true senseless. The intelligent exercise mode selection.
huami-PAI™: The system is based on the user’s heart rate data, combined with the time of day of activity and the human body’s ability to perform a variety of activities. Multi-dimensional physiological data is converted into more intuitive PAI values through algorithms to help users intelligently grasp the amount of daily exercise. In addition, PAI can be tailored to different users, combining age, gender, resting heart rate and other personalized physiological data, to create for each user An absolutely personalized health assessment system.
The system also incorporates a time dimension to analyze physical activity. Users are free to set exercise goals based on their work and rest habits. To reach a PAI value of 100, you can choose to exercise at low intensity for multiple days or focus on high-intensity exercise for several days.
According to the HUNT Fitness Study, maintaining a PAI value of 100 or higher. It is beneficial in reducing the risk of death from cardiovascular disease and increasing life expectancy. The study was led by Professor Ulrik Wisloff of the Norwegian University of Science and Technology, Faculty of Medicine, and involved more than 230,000 participants over a period of more than 35 years. Wisloff), took more than 35 years to complete and involved more than 230,000 participants.
According to Huang Wang, Huami already has a number of wearable devices that support huami-PAI™, including the newly released Xiaomi Bracelets 5. There will be more devices on board in the future, allowing each user to scientifically set and accomplish exercise goals and improve the Physical Health.