Anura: a comprehensive health snapshot from a 30-second selfie
We now reach Nuralogix as we continue our coverage of the bizarre and wonderful fitness and health gadgets at CES 2023. Anura is a wellness app created by this Toronto-based firm. It analyses a 30-second selfie video with artificial intelligence (AI) and generates a variety of health diagnostics.
These contain more than 30 different vitals indicators, such as blood pressure and heart rate. Indicators of physical and mental health include body mass index, skin age, risk scores for illnesses like hypertension and heart disease, along with biomarkers such as blood sugar levels. Mental health readings include stress and depression levels. An impressive list, indeed.
Essential reading: Best fitness trackers and health gadgets
The programme analyses the blood flow in the face and extracts information about the user’s health using a method known as transdermal optical imaging (TOI). Your body’s physiological state, according to Nuralogix, alters the blood flow to your face. Additionally, they found that any device’s camera can be used to detect these changes.
“Light and its respective wavelengths are reflected at different layers below the skin and can be used to reveal blood flow information in the human face.” the company writes on its website.
The system is broken down into five steps
Here are the five steps.
- Video capture: To begin, the TOI system uses any video camera, such as those found in smartphones or tablets, to take a 30 second selfie of the user.
- Face detection and tracking: TOI recognises important areas of interest by automatically detecting and tracking the user’s face.
- Extraction of blood flow data: Because human skin is translucent, it reflects light and its various wavelengths at various layers beneath the skin, providing data on blood flow. Conventional video images contain this data, which TOI captures and securely uploads to the cloud.
- Signal processing and deep learning: To estimate physiological and psychological effects, the DeepAffex engine uses advanced signal processing and deep learning AI models on the facial blood flow data. The business claims that information from about 35,000 distinct users was used to train the system.
- Results: The results from DeepAffex are then sent back to the user’s device for display and further analysis.
Is it accurate?
As you’d expect, some of these vitals stats recordings are more accurate than others. But, Nuralogics, for example, asserts that the standard deviation error for its blood pressure measurements is less than 8 mmHg. Additionally, the data the app generates complies with data protection laws like HIPAA and GDPR. The business is currently looking for FSA approval.
At this time, Anura’s findings are meant to provide direction, and further information for enquiries. Theinformation from the app might be utilised in a more proactive matter, only after it receives FDA approval.
If you are not at CES in Las Vegas, give the Lite version of the application (on iOS and Android) a spin. Here you’ll get some, but not all, of the above listed diagnostics.
I tried the iOS version and it does present some interesting insights. Hey, it judged my appearance to be 10 years younger than my actual age. And I’m not one to argue with scientific findings!
As far as the other data, it’s a decent approximation. But don’t expect accurate smartwatch-type data. For example, while my heart rate data was where I expected it would be, my HRV readings were clearly too high.
Here’s a list of stats you’ll get from the Lite app: heart rate, irregular heartbeats, breathing, heart rate variability, Mental Stress Index, Body Mass Index, Facial Skin Age, waist-to-height ratio, Body Shape Index, General Wellness Score.
You do need to enter your age, height and weight in the app when first opening the account. It would have been more impressive if it could estimate all of the above without that information.
Lots of potential use cases
This technique was initially meant to assist with remote health diagnostics. Not surprising in light of the COVID pandemic. But one could easily imagine other use cases.
One potential application of DeepAffex technology could be to improve or possibly even replace conventional lie detector tests. After all, it has been established that they can be untrustworthy and are known to sometimes produce false positive results. It may be feasible to more precisely ascertain a person’s sincerity by using DeepAffex’s AI to examine facial expressions and emotions.
In fact, the tech could be applied in any circumstances where understanding a person’s emotions is important. By recognising when a consumer is upset it may, for instance, help firms provide better customer service. The possibilities really are endless.
Of course, this does raise the inevitable privacy concerns. If the technology is used in public settings, such as retail stores or transportation systems, individuals may feel that their privacy is being invaded. After all, they would be monitored and analyzed without explicit consent. I mean, it’s not like that sort of thing happens already…
As with most technology, the good comes with the bad! But this is the direction the world is heading in. And there’s no stopping technological progress.
Like this article? Subscribe to our monthly newsletter and never miss out!
