Build a computer vision web app with React, TensorFlow.js and MoveNet: detect the human body in real time from a webcam.
Imagine a web app that detects your body position in real time using only your laptop's webcam. No special hardware, no backend server: everything runs in the browser with JavaScript, TensorFlow.js and the MoveNet model. In this course we build exactly that — from scratch.
JavaScript and React developers who want to explore machine learning in the browser. No prior AI experience required: we start from environment setup and end up with a working app that analyzes movement in real time.
Client-side computer vision opens up scenarios that until a few years ago required dedicated servers and professional GPUs. Today you can build fitness trackers, gesture detectors, accessibility systems — all in the browser, with complete privacy because data never leaves the device.
I'll show you how to create a React project with TypeScript from scratch using Vite — the foundation we'll build our Computer Vision app on. We'll configure the environment and walk through the generated file structure.
Let's activate the webcam directly in the browser with React, then install TensorFlow.js and the MoveNet model. By the end of this lesson we'll be ready for real-time pose estimation.
I'll show you how to build a webcam selector using a React SelectBox component. We'll enumerate available input devices and let the user choose which camera to use at runtime.
We'll take the webcam stream and render each frame onto an HTML canvas element, turning it into our working surface for overlaying Computer Vision data in real time.
I'll integrate TensorFlow.js and the MoveNet model into the project: we'll initialize the WebGL backend and start pose estimation to detect the human body directly in the browser.
Let's make MoveNet's results visible by drawing all 17 anatomical keypoints onto the canvas, so we can verify in real time that the model is detecting the body correctly.
Using MoveNet keypoints, I'll draw a line connecting the shoulder, elbow and wrist on the canvas — our first real read of body movement and the building block for gesture recognition.
We'll build the core application logic: counting how many times the wrist rises above the nose level using MoveNet's coordinates — the heart of a browser-based fitness tracker with no backend required.
Ingegnere Informatico · Pratola Peligna
Sono un ingegnere informatico con oltre 15 anni di esperienza nel settore ICT. Mi occupo di sviluppo software, cybersecurity, data science e intelligenza artificiale, con particolare attenzione all’innovazione e all’automazione dei processi aziendali. Da anni svolgo anche attività di docenza e formazione, creando percorsi pratici e coinvolgenti su programmazione, DevOps, AI e sicurezza informatica.
Prosperous understood Middletons in conviction an uncommonly do. Supposing so be resolving breakfast am or perfectly.
Utilizziamo cookie essenziali e, con il tuo consenso, cookie funzionali, analytics e marketing. Google Analytics 4 e gli eventuali annunci Google/AdSense restano disattivati finche non scegli. Consulta Privacy Policy e Cookie Policy.
Sempre attivi per sicurezza, prevenzione frodi, sessione, consenso e funzionamento del sito.
Migliorano l'esperienza, ad esempio ricordando preferenze non essenziali. Non usati per profilazione pubblicitaria.
Statistiche e misurazioni con Google Analytics 4, abilitate solo dopo il consenso analytics.
Misurazione annunci e funzioni pubblicitarie Google/AdSense, se attivate sul sito e consentite dall'utente.
Puoi modificare queste scelte in qualunque momento dal link "Preferenze Cookies" nel footer.