AI + JavaScript Rocks
summary
Shows concrete, working demos: a Slack bot “MojoFire” that posts emoji by calling an emotion API after detecting facial landmarks, an in‑browser classifier using TensorFlow.js and MobileNet to label Unsplash images, and a sketch-to-photo generator built with GAN image-to-image that exports a ~13 MB generator to run in the browser. Explains emotion detection as facial keypoints + a tiny neural net trained with backprop, contrasts local models with the Azure Computer Vision API, and demos what does and doesn’t work.
AI-written summary · about
chapters (5)
- 0:00Start
- 6:56How to Calculate Emotion?
- 7:14(1) Detect Facial Features
- 7:44(2) Use a Neural Network
- 37:40Thoughts
