YAML to EMT: Build an EMT 3 DL Streamer Image with OS Image Composer (PCB Anomaly + Face Detection)

[rpandya28]

🚀 Overview

This discussion summarizes a demo showing how Intel® OS Image Composer can build an EMT 3 bootable image for edge AI, then run practical DL Streamer workloads using containerized deployment and GStreamer pipelines.

✅ Why OS Image Composer

• 📄 Produce consistent EMT images from a single YAML definition
• 🧩 Reduce post-install steps by baking key dependencies into the image
• 🔁 Improve repeatability for demos, validation, and production rollouts
• 🧾 Make OS builds traceable and easy to regenerate

🧱 What you build (the OS image)

• 🖥️ OS: Edge Microvisor Toolkit (EMT 3)
• 🎯 Target: X86_64 (as referenced in the demo)
• ⚙️ YAML defines: additional package repos, disk artifact type/size, GPT partitions, system config, and boot settings
• 💿 Image output: bootable raw disk artifact

🤖 What you run on top (workloads)

• 🏭 Edge AI Suites: PCB anomaly detection (Docker Compose deployment)
• 🎥 DL Streamer + GStreamer: Face detection real-time pipeline

🔗 Resources

• OS Image Composer: https://github.com/open-edge-platform/os-image-composer
• DL Streamer: https://github.com/open-edge-platform/dlstreamer
• Open Edge Platform Playlist: https://www.youtube.com/playlist?list=PLg-UKERBljNxIMUUPWHpA8ZJ_YUU5tPP9