AI JavaScript Rocks
summary
Demos show practical JavaScript AI: an emoji-overlay bot that detects faces and returns emotion scores using Azure Face API emotion + face coords (originally a Twitter bot, reworked as a Slack bot); browser-only image classification using TensorFlow.js (no native deps) and a MobileNet model that outputs percent confidences (examples show low scores like 9.46% for “terrier”); image-to-image and vid2vid work via pix2pix / GANs (generator + discriminator), exported to TF.js and run in-browser, producing convincing but imperfect generated images.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 3:40MojoFire
- 5:55Vision Libraries
- 6:30Neural Networks
- 13:50Tensorflow Mobile Net
- 14:45CodePen
- 15:55Tensorflow
- 17:25Machine Learning in the Browser
- 19:25Computer Vision Library
- 20:10Accessibility Demo
- 21:20Stuffed Animals
- 22:30Image to Image
- 23:30generative adversarial
- 26:30vidtovid
- 27:25segmented image
- 28:05bird image
- 29:50form gam
- 31:20summary
- 32:00find out more
- 32:25final thoughts
