Project
Check out my machine learning project! I used ML5, and Posenet built on tensorflow to train and classify poses . You’ll find demos of my creations and details on the frameworks and datasets I used. Get ready to be impressed, or at least mildly amused.
Must be logged in for security purposes. please contact me if you want to try it
Goal
Achieve Real-time Pose Estimation from browser
This is a simple application where the user’s image would be captured by a camera and processed by PoseNet to identify the location and orientation of key body parts. The positions of these body parts would be compared to a pre-trained model for each of the four poses using a distance-based comparison or a machine learning approach. Once the pose that best matches the user’s body position has been determined, the application will highlight the pose performed. With the use of P5.js , a canvas overlay can be used to map the key point to a webcam video and acheive realtime pose estimation. With careful tuning and testing, such an application could provide a fun and interactive way for users to practice and improve their yoga or exercise routines.
Machine-Learning Frameworks
ml5.js | TenSorFlow.js
ml5.js is a JavaScript library that provides a simple and accessible way to use machine learning models in web-based projects. it is built on top of TensorFlow
JS Librabries
p5.js
p5.js is a JavaScript library that simplifies the creation of interactive graphics and creative coding projects for the web.