aditya2023003/hofteamram

correct your Yoga posture using machine learning

# Team Ram correct your Yoga posture using machine learning

introduction:

Yoga is a popular form of exercise and relaxation practiced by millions worldwide. Detecting yoga poses automatically using machine learning can be beneficial for practitioners to ensure correct posture and alignment. In this project, we aim to develop a web application that uses machine learning techniques to detect and classify yoga poses in real-time using JavaScript.

Technologies Used:

TensorFlow.js: A JavaScript library for training and deploying machine learning models in the browser and on Node.js.

PoseNet: A pre-trained deep learning model that can estimate human pose in real-time with a single camera input.

HTML/CSS: For creating the user interface of the web application.

JavaScript: For implementing the machine learning model, handling real-time video input, and displaying the results.

Project Steps:

Data Collection:

Gather a dataset of images containing various yoga poses. Ensure diversity in yoga poses, backgrounds, lighting conditions, and body types. Annotate the dataset with labels corresponding to different yoga poses. Model Training:

Preprocess the dataset and split it into training and testing sets.

Use TensorFlow.js to build and train a machine learning model for pose detection. PoseNet is a popular choice for this task. Fine-tune the model as necessary to improve accuracy. Web Application Development:

Create a user-friendly interface using HTML/CSS to display the webcam feed and detected yoga poses. Integrate the trained machine learning model using TensorFlow.js to perform real-time pose detection. Implement functionality to capture video input from the user's webcam. Real-Time Pose Detection:

Utilize the webcam feed to capture real-time video input. Pass each frame of the video through the trained PoseNet model to detect and classify yoga poses. Display the detected poses overlaid on the video feed in real-time. User Interaction:

Provide feedback to the user regarding the correctness of their yoga pose alignment. Highlight the cautions and benefits Offer suggestions for adjustments to improve posture if necessary.

Testing and Evaluation: Evaluate the performance of the model on the testing dataset. Conduct user testing to gather feedback on the accuracy and usability of the web application. Make improvements based on user feedback and performance evaluation results.

Conclusion: In this project, we have demonstrated the development of a web application using machine learning and JavaScript to detect and classify yoga poses in real-time. This application can be a valuable tool for yoga practitioners to improve their posture and alignment, leading to a more effective and safe practice. With further refinement and optimization, such technology has the potential to enhance the yoga experience for practitioners of all levels.

Video Demo Link : https://www.youtube.com/watch?v=pIDRLUfNZUE

Contact details: [email protected]