Robust Intelligent Posture Estimation for an AI Gym Trainer using Mediapipe and OpenCV

Abstract

Robust Intelligent Posture Estimation is an important aspect of an AI Gym Trainer that can help fitness enthusiasts improve their workout technique and prevent injuries. This research presents an approach to achieve accurate posture estimation using Mediapipe and OpenCV. Mediapipe is a machine learning framework that provides pre-trained models for human posture estimation, while OpenCV is a popular computer vision library that offers a range of functions for image and video processing. The proposed solution integrates the strengths of both tools to develop a robust posture estimation system. The system first captures the user’s video feed and passes it through MediaPipe to detect the human body landmarks, then, OpenCV is used to calculate the angles between the detected landmarks in order to analyze the posture. The system provides real-time feedback to the user on their posture and suggests reparative measures. The use case that has been used for this research was repetitions for bicep curls. The proposed system can be tested on various exercises, such as squats, push-ups, and lunges. It can accurately estimate the posture of the user in different lighting conditions and is robust to occlusions and background clutter. The system can be deployed as an AI Gym Trainer and can help fitness enthusiasts improve their form and technique while reducing the risk of injury.