Browser-based virtual try-on system using Pose Estimation (ml5.js / BlazePose) and Body Segmentation (TensorFlow BodyPix) for real-time accessory and clothing overlay.
-
Updated
Apr 20, 2026 - JavaScript
Browser-based virtual try-on system using Pose Estimation (ml5.js / BlazePose) and Body Segmentation (TensorFlow BodyPix) for real-time accessory and clothing overlay.
A Computer Vision web app that sees you, hears you.
This project is a real-time face landmark detection web application built using React and TensorFlow.js. It uses the FaceMesh model to detect 468 facial landmarks from a live webcam feed and renders a triangulated mesh overlay on the face using the HTML Canvas API. The goal of this project is to demonstrate how machine learning models
To associate your repository with the tensflowjs topic, visit your repo's landing page and select "manage topics."