Product Recognition in Retail/E-Commerce with Azure AI
summary
A live demo builds a small web app that uploads photos, calls Azure AI services (CustomVision object detection), and redirects to a product page when a prediction crosses a threshold. Training uses a minimum of ~15 images per tag (author advises 30–50 for robustness); quick training yields 91.7% precision, 100% recall on four product classes. The demo shows endpoint keys, JSON prediction scores, practical failures when stock is absent, and a working prototype that reliably detects closeup product images.
AI-written summary · about
chapters (9)
- 0:00AI in retail and e-commerce
- 4:40Live demo: product recognition
- 8:11Azure AI resource setup
- 12:52AI building blocks overview
- 15:23Training custom vision model
- 22:41Testing model and API
- 25:57Demo app with AI
- 31:35AI use cases in e-commerce
- 34:27Q&A
related talks
42mProduct Recognition in Retail/E-Commerce with Azure AI
01h 04mBuilding and Training your own Custom Image Recognition AI
58mPlaying the game with Azure Cognitive Services
01h 00m5 Levels of AI in Microsoft Azure
57mWhat's that fruit?
41mAI + Machine Learning for .NET Developers
01h 03mEasier AI allows everyone to build smart apps
16mHow to get started with Computer Vision AutoML in Azure