Three-dimensional monitoring of weightlifting for computer assisted training (original) (raw)
Related papers
Weight Training Analysis of Sportsmen with Kinect Bioinformatics for Form Improvement
ArXiv, 2020
Sports franchises invest a lot in training their athletes. use of latest technology for this purpose is also very common. We propose a system of capturing motion of athletes during weight training and analyzing that data to find out any shortcomings and imperfections. Our system uses Kinect depth image to compute different parameters of athlete's selected joints. These parameters are passed through certain algorithms to process them and formulate results on their basis. Some parameters like range of motion, speed and balance can be analyzed in real time. But for comparison to be performed between motions, data is first recorded and stored and then processed for accurate results. Our results depict that this system can be easily deployed and implemented to provide a very valuable insight to dynamics of a work out and help an athlete in improving his form.
Multiple Kinect based system to monitor and analyze key performance indicators of physical training
Human-centric Computing and Information Sciences
Using a single Kinect device for human skeleton tracking and motion tracking lacks of reliability required in sports medicine and rehabilitation domains. Human joints reconstructed from non-standard poses such as squatting, sitting and lying are asymmetric and have unnatural lengths while their recognition error exceeds the error of recognizing standard poses. In order to achieve higher accuracy and usability for practical smart health applications we propose a practical solution for human skeleton tracking and analysis that performs the fusion of skeletal data from three Kinect devices to provide a complete 3D spatial coverage of a subject. The paper describes a novel data fusion algorithm using algebraic operations in vector space, the deployment of the system using three Kinect units, provides analysis of dynamic characteristics (position of joints, speed of movement, functional working envelope, body asymmetry and the rate of fatigue) of human motion during physical exercising, ...
A Kinect-based rehabilitation exercise monitoring and guidance system
2014 IEEE 5th International Conference on Software Engineering and Service Science, 2014
In this paper, we describe the design and implementation of a Kinect-based system for rehabilitation exercises monitoring and guidance. We choose to use the Unity framework to implement our system because it enables us to use virtual reality techniques to demonstrate detailed movements to the patient, and to facilitate examination of the quality and quantity of the patient sessions by the clinician. The avatar-based rendering of motion also preserves the privacy of the patients, which is essential for healthcare systems. The key contribution of our research is a rule-based approach to realtime exercise quality assessment and feedback. We developed a set of basic rule elements that can be used to express the correctness rules for common rehabilitation exercises.
Computer-Supported 3D Kinematic Patterns Capture and Analysis in Handball
2018
Gross human movement biomechanics resides on a well-known inverse dynamics paradigm, established for the first time in times when high-speed photography served as a motion capture technique (Braune & Fischer 1895). Since then, the high-speed photography method has successfully served for years to accurately record and accumulate kinematic data that were subsequently either directly interpreted or processed. The 20th century, being the age of electronics, brought the first technical solutions to automatically detect characteristic points (markers) in a TV picture frame (in 1967, transition from black to white detection principle) as the starting point for gathering kinematic data (Furnee 1989). This marks the beginning of the development of a plethora of commercial technical solutions to realize accurate and rapid (semi)automatic 3D kinematic data capture (Medved 2001). The beginnings of our Biomechanics Laboratory at the Faculty of Kinesiology, University of Zagreb, were also marked...
A Motion Capture System for Sport Performance Analysis Based on Inexpensive RGB-D Sensors
2019
Motion capture for gesture and performance analysis in sports usually requires tracking of several anatomical landmarks with high spatial accuracy and sampling frequency. In addition, a model of the outer surface of the body can also be useful, e.g. for aerodynamic studies. In this context, we developed a motion capture system based on four inexpensive commercial RGB-D cameras. Our system produces both skeletal poses and 3D meshes of moving human bodies at high frequency with good accuracy.
Muscle Vision: Real Time Keypoint Based Pose Classification of Physical Exercises
ArXiv, 2022
Recent advances in machine learning technology have enabled highly portable and performant models for many common tasks, especially in image recognition. One emerging field, 3D human pose recognition extrapolated from video, has now advanced to the point of enabling real-time software applications with robust enough output to support downstream machine learning tasks. In this work we propose a new machine learning pipeline and web interface that performs human pose recognition on a live video feed to detect when common exercises are performed and classify them accordingly. This exercise pose classification pipeline experimentally determines, in a real-time mobile environment, the type of physical exercise being performed from a predefined set of examples. In contrast to existing models that use image recognition directly, our classification model strictly focuses on three dimensional coordinates for key landmarks on the human body, which requires a considerably smaller input size an...
Kinect Joints Correction Using Optical Flow for Weightlifting Videos
2015 Seventh International Conference on Computational Intelligence, Modelling and Simulation (CIMSim), 2015
To ease a coach in weightlifting training, automatic weightlifting pattern evaluation is required. In order to do that, the motion tracking process is always needed. Kinect sensor is one of the popular sensors for that. However, there is a problem with skeleton created by the Kinect sensor because of self-occlusion. Hence, in this paper, we develop a joint correction process for 3 types of joints including hands, feet, and knees since these joints are sometimes provided incorrectly. However, we only correct these joints in the "the first pull to the transition from the first to the second pull" (first-step) in snatch, and clean and jerk weightlifting and "the turnover under the barbell to the catch phase" (second-step) in clean and jerk weightlifting. This is because miscalculation occurs only in these steps. We utilized fast cross-correlation and the Lucas-Kanade algorithm to compute the optical flow of the consecutive frames. From that, we then correct the joints if they are misplaced from the predicted joints. Our system provides better joints and more preferable to human eyes than the original Kinect skeleton.
Autonomous sports training from visual cues
2003
Computer driven biometric analysis of athlete's movements have proven themselves as effective sports training tools. Most current systems rely on the use of retro-reflective markers or magnetic sensors to capture the motion of the athlete, so the biometric analysis can be performed. Video based training tools have also proved to be valuable instructional aids, however most require significant human interaction for analysis to be performed. This paper outlines an ongoing project focussed on capturing posture without the use of any markers or sensors, while still capturing enough information for an automated analysis to be performed. The approach taken to solving this problem is presented, as well as the current state of development of a an instructional aid for golfers.
Exercise plays an important role in our day to day life as it helps people remain in shape, fit and to prevent from many disease. Regular physical activities such as weight training and cardio exercises are part of everyday modern life. If performed correctly, it contributes to the health of a person. Exercises helps to prevent obesity and stimulate the immune system. Many people practice physical exercises without an assistance of an expert in home. This paper aims to present a software that offers virtual trainer with real-time feedback and the assessment score to different exercise postures presented by an animated 3D character using Kinect sensor. This tool allows people to observe the correct execution of each exercise. Recognition of the exercise has been performed using Random Forest (RF) classifier. The computer must first understand what a user is doing before it can respond. This has always been an active research field in computer vision, but it has proven formidably difficult with video cameras. With help of Kinect sensor, the computer directly sense the third dimension, making the task much easier