Non-traditional usage of Cloud for Science Art
summary
Uses Azure Functions + Face API to build an interactive exhibit that captures a photo, extracts ~29 facial landmarks, aligns images (three-point affine), and averages the last 10 frames to produce a blended portrait; client runs as a UWP app (or Raspberry Pi) with local face detection to avoid excess cloud calls. Back end exposes URLs, experiences a cold-start warmup, ran in a Moscow museum and via a Telegram bot (peopleblenderbot) and Azure ML trains GANs (11-hour runs) for fully generated art.
AI-written summary · about
chapters (7)
- 0:00Introduction to cloud art
- 2:31Cognitive portrait technique
- 7:49Interactive museum exhibit
- 13:46Azure Functions for processing
- 20:52Extending exhibit online
- 26:51AI-generated drawings
- 30:23Summary and resources
related talks
50mScience Art in the Cloud: How Microsoft Azure helped me to Express Myself
40mScience Art in the Cloud: How I built a Museum Exhibit
32mSelf-Portraits in JavaScript
39mAI + JavaScript Rocks
35mHow to AI in JS?
35mAI JavaScript Rocks
48mThe Future of Machine Learning & JavaScript
33mApplied AI: Real-world use cases for Microsoft's Azure Cognitive Services