Research scholar at IITGN develops device to overcome fitness monitoring limitations

The next international conference of IMWUT is scheduled in Melbourne in October where a presentation on JoulesEye will be held, he adds.

Published by the Association for Computing Machinery, IMWUT is a journal covering the technologies, fields, and categories related to computer networks and communications, hardware and architecture, and human-computer interaction. The Association for Computing Machinery is the world’s largest educational and scientific computing society.

Giving a background on the research project, Adhikary, a research scholar at IITGN’s Computer Science Engineering department, shares with The Indian Express, “People regularly use their phones and watches to keep track of their physical activity and fitness levels. Two of the most popular measures the users look at while exercising are heart rate and calories. The phones and watches have become quite good at measuring heart rate but calories are widely inaccurate.”

“One of the most accurate ways of getting calorie expenditure is using an indirect calorimeter. But they are very expensive, bulky to use, and almost impossible to carry around. Another measure that users find useful is respiration. But except a coarse estimate while at rest, respiration signal is almost entirely missing from consumer devices,’ he stresses.

“The calorie burn (or energy expenditure estimates) of commercial smartwatches are erroneous by over 40 per cent. Our study corroborated these findings. We conducted trials for our system JoulesEye on 54 participants ( (aged between 25 and 54) who performed cycling and running activities. We compared the performance of our system with a gold standard indirect calorimeter and found that the error of our system is only 5.8 per cent,” Adhikary, who joined PhD at IITGN in 2019, further explains. He completed his BTech from Gauhati University in Assam.

The trial study was conducted in the SMASH lab of Carnegie Mellon University (CMU) in Pittsburgh, the United States, during Adhikary’s stint as a Fulbright scholar between August 2022 and May 2023 where Prof Mayank Goel from the CMU and Prof Nipun Batra from IIT Gandhinagar were his advisors for this project.

The findings revealed that an ultra-low-resolution thermal camera that is small enough to fit inside a watch or other wearables is sufficient for accurate calorie burn estimation. The results suggested that JoulesEye is a promising new method for accurate and reliable calorie burn estimation, Adhikary adds.

JouleEye is a thermal imaging-based approach to measure respiration and caloric expenditure. It uses a thermal camera to observe temperature changes in the nostrils. Signal processing and AI are used to extract respiration and caloric expenditure from the temperature change, he explains.

A smartwatch prototype for JoulesEye has been developed and a low-resolution thermal camera has been engineered keeping in view cost, size, and battery.

“Even with very low-resolution thermal imagery, we can still extract respiration rate from temperature change in nostrils. Currently, the user needs to look for over 40 seconds into the watch. We are working to reduce this time further,” the research scholar sums up.