Build a look alike engine with machine learning and Angular
summary
An Angular app uses Azure Custom Vision plus the Bing Image Search API and Azure Functions to build a doppelgänger classifier, uploading image URLs and training models programmatically instead of manual uploads. Model needs at least five images per tag, precision/recall are shown and a test returns a spurious 99.9% match, so iteration matters. The pipeline automates collecting ~150 images (30 tags × 5 images) and posts them to the training endpoint; the demo runs but requires more data to be reliable.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:08The Doppelganger Game
- 2:05Tech Talk Bingo
- 5:16The Machine Learning Algorithm
- 9:20Add Images
- 10:55Train the Algorithm
- 15:46Custom Vision Service
related talks
26mLet's Build a Doppelganger Game!
01h 04mBuilding and Training your own Custom Image Recognition AI
44mHop on the serverless adventure
46mCreate Smart Angular Apps Using Azure Computer Vision
01h 00m5 Levels of AI in Microsoft Azure
42mProduct Recognition in Retail/E-Commerce with Azure AI
42mProduct Recognition in Retail/E-Commerce with Azure AI
01h 01mGetting Started with Computer Vision