Deep learning enhanced smart mirrors can guide fitness

Researchers at the University of Bresia in Italy recently developed a computer vision system for smart mirrors that can improve the efficiency of fitness training in the family and gym environment. This system is introduced in a paper published at the meeting of the International Sports Biomechanics Society. The system is based on a deep learning algorithm and can identify the human posture recorded in fitness videos after training.

This low -cost computer vision system uses the osteoma algorithm to run on the embedded Nvidia Jetson Nano device with two fish -eye cameras. The system can extract information from the human sports video clips taken by artificial intelligence (AI) algorithms. The paper proves the accuracy of the arm movement in the measurement of simple fitness exercises (such as the biomotropic roll).

Researchers said that the advantage of the system is that there is no need to contact users. Using cameras and AI applications, they can understand and evaluate physical exercise, detect posture errors, and analyze simple fitness exercises. Now, the system analyzes a simple body variable (such as elbow angle and hand position), and they are still trying to improve AI's evaluation capabilities.

Researchers are working with their business partners AB-Horizon to design a smart mirror that can evaluate fitness exercises like human private coaches, or evaluate in a more comprehensive way. For example, it can calculate the number of specific exercises made by the user, and at the same time, it can also detect basic exercise in different parts of the body, such as traction, flexion, and rotation.

All fitness -related information from mirror detection and calculation will be displayed on the mirror and changes in real time, so that users can understand their own situation in real time during the exercise process, make scientific adjustments, thereby improving the training scores.

The researchers evaluated the system through a series of tests, especially when the user conducted biceps exercise, the system could track and make health predictions.

Researchers have found that through carefully designed and calibrated software, their low -cost visual systems can provide valuable fitness -related data during simple fitness exercises. After integrating it into a smart mirror, the system can significantly help users who train in family and gym without guiding coaches.

- END -