Hello world!
May 25, 2022
1 Comment
Welcome to WordPress. This is your first post. Edit or delete it, then start writing!
Full Body Tracking is a feature that helps FitVR user to project their real body movements into an avatar and get an immersive experience in VR while working out. This is important because the sense of body ownership is defined as an illusion when users perceive the virtual body as their own body. For body pose tracking an ML solution like MediaPipe is used which infers Pose Landmark Model that detects 33 landmarks on the user’s body. Using camera vision as an input source, the body is tracked and aligned with the HMD position to get better results. As we do not need any external sensors, this feature is highly adaptable among users. Using full body tracking as a baseline we can implement many features like gamification, customized workouts, and hand tracking which enhances the user experience. Also, full body tracking is crucial to ensure that users perform desired physical activities correctly, either to improve health outcomes or to lower the risk of injury. There are many systems in the market which track the full body and transform it into an avatar with the help of different sensors/markers and tracking devices, but we constructed a system with minimal resources which use just an ML algorithm and a system. As part of this feature, we accomplished creating a sense of control and embodiment for the user in a virtual environment while performing a workout.
FIT-SWEAT is an exergame where the users get to perform exercises through an interesting game that has a scoring system that keeps them motivated. If the regular workouts aren’t motivating for some, then this application will keep them playing and working out. It includes a demo game where the users can get used to the environment and the game itself. And then, the actual game can either be played in a single-player or a multiplayer mode to collaborate with friends in the same room. All the users must do is to keep punching the cubes that keep spawning towards them and dodge the dodge obstacles. The health is initially at a hundred and it keeps reducing when a dodge obstacle is touched. To increase health, the red balls should be punched. The game becomes more interesting and challenging with the color brain game. There are two cubes with different colors along with a color word written. The users have to hit the cube that has the color of the text. To make the users utilize the space around and increase their movement, they also get to play the Whack-a-mole mode. There are colorful circles that appear on the floor that needs to be stomped onto to earn bonus points and health. There is a streak system when the cubes are hit continuously. The users will be enthralled to beat the high score. In the multiplayer mode, both the players will have to coordinate with each other for the color brain game and the whack-a-mole mode. They must make sure to perform their best together. All of this with the heartrate system that will control the speed of the game and the Dynamic Music System that will add compelling background music to the game, all included within our attractive environment .
In the customized workout feature, we provide the user with a guidance system based on the FastDTW algorithm which performs posture detection and correction. This algorithm uses a multilevel approach that recursively projects a solution from a coarse resolution and refines the projected solution. In the system, there are Trainer and Trainee modes where we train the data initially and compare it with the real-time data to find the accuracy with which the user performs a workout. Along with the accuracy, the user will be given different forms of feedback like matching the stick figure which is required to correct his posture while working out. Features like bone highlighter and graphical representations are implemented to provide analysis of his body movements during his workout. As part of the bone highlighter, we allow users to select the set of muscles to be trained and customize their workout plan. The Graph feature is implemented for users to get comparison details and what corrections are required to achieve high accuracy. The side and front view avatars will be available throughout the workout to see how the trainer performs the workout without any wrong postures. Users can plan and customize their workout routine with our efficient tracking and guidance system.
With the User Profile Management system, the users start by logging in to the application by filling out a few demographic details like height, weight, BMI, age, and gender. This data will then be used later for the statistics and history, where the calories burnt, and average heart rate are calculated. These demographic details can also be edited later from the Edit profile option on the profile after logging in. All the details of the users’ profiles are stored locally in an SQLite database. After the users log in, they have the option to choose to do the regular workouts or to head to play FIT SWEAT. The score they make in FIT SWEAT will also be stored in the database along with their average heart rate. The accuracy values and data points from customized workout suggestions are also saved to this dabatase. The users can always come back to their profile from any mode.
Monitoring our progress while exercising is crucial for improving our performance. In the real world, we receive feedback from our trainers or the training apps. In our FIT-VR application, we have built the “Statistics, and History” function, which is in charge of giving the user feedback following each session. It primarily consists of two sections: Daily statistics, which are provided following each workout, and History, which displays the user’s progress over the previous seven days of workouts. The major metrics that should be kept track of during the workout are the number of sets performed, the calories burned, and the maximum and average heart rates. In our application, the user increments the number of sets after completing a set of repetitions, and the heart rate is recorded using the “Polar H10” heart rate sensor, which is a sensor wrapped around the chest, and the calories burned are computed using the user’s demographic data such as height and weight, as well as exercise specific data such as MET (Metabolic Equivalents) value and the number of sets completed. A dashboard including all of this data is created after analysis and given to the user. Users can utilize this function to plan ahead for their upcoming workout sessions.
Welcome to WordPress. This is your first post. Edit or delete it, then start writing!
Sem felis dictumst luctus mattis laoreet fringilla placerat eleifend senectus. Aliquam nulla urna commodo cubilia magnis orci viverra amet quisque ultrices vehicula. Lobortis tellus sed tortor mi imperdiet odio etiam.