A team of researchers from Italy have used a deep learning algorithm and developed a smart mirror that is capable of providing effective fitness training
Physical fitness is essential to enhance your health and reduce the risk of developing several life-threatening diseases such as type 2 diabetes, PCOD, cancer, and cardiovascular disease. Due to busy work schedules and the huge fees of a personal trainer, people neglect their health. Hence, researchers at the University of Brescia in Italy invented a low-cost computer vision system using a deep learning algorithm for a smart mirror trained to recognize human gestures in video recordings which would enhance the effectiveness of fitness both in home and gym environments. All the fitness-related data traced and calculated will be visible on the mirror and will change in real-time enabling users to keep track and improve their training performance.
“A vision system, like the one we developed, can extract information from images by means of an AI algorithm,” says Bernardo Lanza, one of the researchers who conducted the study. “Our most recent paper demonstrates the accuracy of our system in measuring arm movements in simple fitness exercises, such as biceps curls. The main advantage of our system is the absence of objects in contact with the user. With cameras and AI applications, we understand and assess body motion, detect postural errors, and analyze simple fitness exercises. Nowadays our system analysis is based on simple body variables (elbow angle, hand position…) but we are working to improve the evaluation capability of the machine.” says Lanza.
The researchers developed a software design that can be implemented for a comprehensive prototype of the smart fitness mirror envisioned by AB-Horizon (commercial partner). The device uses a skeletonization algorithm (i.e., a deep learning algorithm that can attain skeletons from images) running on an embedded Nvidia Jetson Nano device with two fisheye cameras. The system is trained to detect human movements in video captured by two fisheye cameras. This allows users to track the count of repetitions they performed for a specific exercise and also trace fundamental motion (e.g., traction, flection, rotation, etc.) of different body parts.
“We evaluated the accuracy of the vision system in understanding the different phases of an exercise. For this project, we collaborated with AB-Horizon, our commercial partner. In addition to designing the gym machinery, our partner will integrate the vision system with their prototype. Their experience in the fitness industry allows us to develop our software using athletic principles and a personal trainer from the company also guides us through the testing process. As part of future developments, an intelligent evaluator will detect the exercise.” said Lanza.
Lanza and his colleagues are further developing intelligent evaluators to can interpret qualitative information by analyzing raw body kinematic data. To train the model, they will need large amounts of data during fitness tests with both athletes and less-experienced fitness trainees.
Click for the Published Research Paper and Demo Video
