The technology’s infrared time of flight 3D camera automatically films a person’s bare feet for signs of peripheral oedema (foot swelling), which can present as a symptom of heart failure around two weeks prior to readmittance to hosptial.
The telemonitoring solution has been developed to mitigate against hospitalisation as the two-week window provides enough time to administer medication. According to the company, the solution could prevent up to 75 per cent of hospital readmissions.
Unlike traditional monitoring methods that rely on patient input, this system operates autonomously, capturing foot images multiple times a day without any active participation required from the patient. This data is then sent to the cloud for processing.
Dr Shamus Husheer, CEO and chief scientific officer Dr Oriane Chausiaux told The Engineer that the device is in its third generation, adding that the first version used Raspberry Pi cameras mounted on an aluminium bar.
“The problem with that is that you need a very wide baseline in order to get a certain uncertainty in depth,” said Dr Husheer. “But when people’s feet are close to you, you need a narrow baseline, otherwise you don't actually get any data, whereas time of flight cameras - which was the technology in the Microsoft Kinect for Xbox One cameras - have an approximately linear increase in noise with depth, as opposed to a quadratic increase in noise with depth.”
Related content
Dr Husheer added that the Microsoft Kinect was initially disregarded for 3D data acquisition as it was deemed insufficiently sensitive, but with ‘a bit of fun extra maths’, software development and hardware modifications the team produced a device that allowed them to conduct a small feasibility trial that demonstrated the validity of their measurements.
“We'd proven with the raspberry PI that you could use an optical measure to detect the clinically relevant changes in feet around heart failure,” said Dr Husheer. “But it was impractical to stick those things in people's homes… so we had to make a device that was sufficient to put in up to a couple of hundred patients’ homes.”
The team also needed to acquire real patient data to use as training sets, and to this end the team initially used Blender.
“It's a graphics 3D graphics programme to simulate models in 3D of people's feet and train the neural networks," said Dr Husheer. “Once you got data coming in…you can annotate it and look for the things that are sort of lying outside what your model collects, and keep that going. Now we have some large fraction of a petabyte of such data flying around.”
To date, the system has obtained regulatory clearances, including CE marking under the Medical Device Regulation (MDR) in Europe, and exemption status in the United States, pending further clinical studies.
Dr. Chausiaux and Dr. Husheer emphasised the importance of privacy considerations in deploying the technology, ensuring that patient confidentiality is maintained while delivering clinically meaningful insights. While the current focus is on heart failure monitoring, the potential applications of this technology could extend to other conditions, such as high-risk pregnancies.
Heartfelt Technologies were among the winners of the sixth #21toWatch awards, which acknowledge leading up and coming innovators and entrepreneurs in Cambridge and the East of England.
Five ways to prepare for your first day
If I may add my own personal Tip No. 6 it goes something like this: From time to time a more senior member of staff will start explaining something...