Applied Custom Vision
summary
Shows how to turn existing RTSP camera streams into actionable signals by training Azure Custom Vision models (roughly 25+ labeled images per class), deploying via edge inference on Jetson or Kubernetes, and piping outputs—bounding box, label, timestamp, confidence—into filters and alerts (example: 80% threshold, save video snippets). Practical demos include counting people, detecting brand logos (Palm), reading a thermometer in a server room, GDPR-friendly aggregated counts, and notes on hardware acceleration, 5G and OpenVINO for production.
AI-written summary · about
chapters (9)
- 0:00Applied Custom Vision intro
- 3:00What is Custom Vision
- 8:22Computer vision concepts
- 13:07Build Custom Vision model
- 16:42Edge devices with Azure IoT
- 20:18Deploy and test model
- 23:55Real-world applications
- 26:03Custom Vision tools
- 28:14Q&A
related talks
01h 33mComputer Vision in Azure
46mBuild apps with state-of-the-art computer vision
46mComputer Vision to Secure Your Surroundings with AI/ML Solution Built Using Open Source Tools at the Edge
33mBuilding end-to-end computer vision solutions from deep learning models
46mBuilding computer vision applications Q&A
19mAzure-enabled vision AI with NVIDIA AI Enterprise and Jetson
26mUse computer vision at the edge to improve operations with AWS Panorama
28mVision, Big Data and AI