A team of researchers at Cornell University has created a new type of smart t-shirt that is capable of tracking a person's posture and exercise routine.

Dubbed SeamFit, the garment features flexible conductive threads sewn into the neck, arm and side seams of a standard short-sleeved t-shirt. The shirt’s developers explained that the wearer does not need to manually log their workouts thanks to an artificial intelligence (AI) pipeline that detects movements, identifies the exercise and counts reps. Once the wearer is finished with the shirt, the wearer removes the circuit board at the back neckline and then places the shirt in the washing machine.

Source: Louis DiPietro

Designed potentially for athletes, fitness enthusiasts and patients engaged in physical therapy, SeamFit — unlike existing body-tracking clothing that tends to be tight and restrictive or embedded with bulky sensors — can be treated like traditional clothing with a more comfortable fit, the developers explained.

Making SeamFit appropriate for monitoring fitness is that as the wearer exercises, the threads' capacitance — otherwise known as their ability to store charge — changes as the threads move, deform and interact with the human body. Meanwhile, the circuit board at the back neckline measures the capacitances and then communicates them via a Bluetooth connection to a computer while a lightweight signal-processing and machine-learning pipeline subsequently deciphers those movements.

The team tested the shirts' performance on 15 volunteers who performed 14 different exercises such as lunges, sit-ups and biceps curls, while wearing the SeamFit t-shirt. Once complete, the team determined that SeamFit's model classified the exercises with a 93.4% rate of accuracy while also successfully counting reps, with counts that typically were off by just less than one.

The t-shirt is detailed in the article, “SeamFit: Towards Practical Smart Clothing for Automatic Exercise Logging,” which appears in the journal Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies.