Can a piece of code fix my posture
summary
A desktop app built with Angular, Electron, p5.js, ml5.js uses ml5’s POSENET model to detect body joints, trains on each user’s webcam data (correct and incorrect poses) and flags sustained poor posture with a system notification after about a minute. Demo shows webcam accuracy is good on common webcams, training is per-user for better results, source is open‑source (free to fork but not sell), and a move to MoveNet/TensorFlow models is planned.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker intro
- 1:06What is Angular
- 3:10Key technologies overview
- 5:58ML5.js and PoseNet explained
- 6:59Demo: Posture detection app
- 8:46Future improvements and open source
- 9:50Q&A session begins
- 12:40Maths and learning ML
- 14:29Closing remarks
related talks
28mBuilding Motion Controlled Websites in JavaScript
13mml5.js - Friendly machine learning for the web
19mHand gesture-controlled robot through the Internet
30mBuilding Machine Learning Models on Edge with React
05mIncredible things your community is creating with TensorFlow.js
22mStrike a Pose: Gesture Recognition in JS with Machine Learning
25mHands-free Coding with Gaze Control in JavaScript
26mSelf-Portraits in JavaScript