Metro AI Suite#
Edge AI Suites are collections of open, industry-specific AI software development kits (SDKs), microservices, and sample applications for independent software vendors (ISVs), system integrators, and solution builders. The Metro AI Suite focuses on accelerating development of Edge AI solutions for safety and security, smart city, and transportation.
Achieve High PerformanceSee how to get the most out of your hardware, using Open Edge Platform components.
See what options you have to use Open Edge Platform components with your AI solutions based on the NVIDIA ecosystem.
Learn how to make your AI solutions more secure, on the example of the Smart Intersection Agent & Smart Intersection Agent sample app.
A selection of tutorials that will walk you through creating Metro Vision AI based solutions, from setting up your first application to running inference data processing pipelines.
This section describes the Metro AI Suite workflow for hardware partners.
This section describes the Metro AI Suite workflow for developers.
A comprehensive development environment for computer vision applications using Intel’s optimized tools and frameworks.
A comprehensive development environment for generative AI applications using Intel’s optimized tools and microservices.
A comprehensive demonstration environment for computer vision applications using Intel’s optimized tools and frameworks.
Metro AI Suite Samples#
Visual AI Demo Kit
Visual AI Demo Kit: Demo kit and sample collection for Metro Vision AI use cases.
Loitering Detection: A demo kit sample for detecting and monitoring human presence.
Smart Intersection: A demo kit sample for using AI and multiple camera feeds to improve traffic flow control.
Smart Parking: A demo kit sample for parking management in real-time.
Live Video Analysis
Live Video Alert Agent: An agent for AI-powered real-time alerting for live video streams.
Live Video Captioning: AI-powered analytics and captioning for live RTSP video streams.
Live Video Captioning RAG: RAG-based captioning, used by the Live Video Captioning app.
Live Video Search: A sample that enables semantic search on live Frigate streams.
Real-Time Scene Understanding: AI-powered live video analysis for RTSP streams.
Agentic AI
Agentic Predictive Maintenance for Critical Infrastructure: A sample that demonstrates predictive maintenance using multi-agent reasoning.
Agentic Smart Community: A sample that demonstrates a smart community using multi-agent reasoning.
Enterprise Data Intelligence - Multi-Agent Enterprise Knowledge Assistant: An agent-native automation platform for end-to-end enterprise data intelligence.
Smart Building Digital Twin: A complete smart-building monitoring simulation with end-to-end deployment.
Smart Route Planning Agent: An agent for route optimization that uses multi-agent communication.
Smart Traffic Intersection Agent: An agent for analyzing and managing traffic scenarios at intersections.
Video Search and Discovery
Image-based Video Search: A sample that detects and retrieves objects in large video datasets in real time.
Video Search and Summarization: Search through videos and summarize them.
Visual Search and Q&A: An AI assistant with multi-modal contextual search engine for providing more relevant answers.
Video Infrastructure
Smart NVR: Convert conventional network video recorders into intelligent, context-aware systems.
Video Processing for NVR: Evaluate and optimize video processing workflows for network video recorders.
VMS Adapter Plugin: A plugin for integrating AI Analytics Applications with third-party video management systems.
Sensor Fusion
Sensor Fusion for Traffic Management: A sensor fusion framework for intelligent monitoring of traffic in fog, rain, or darkness.
Software Development#
Metro SDK for application- and framework-level development in Metro AI Suite.
SDK for optimized, low-level, performance-critical video pipeline development on Intel® hardware.
Optimizes and deploys AI models across Intel® hardware in cloud, edge, and on premises environments.
Hosts AI models via networks, e.g. enabling Smart Parking to share a single vehicle-detection model.
Deep Learning Streamer (DL Streamer)
Open-source GStreamer framework-based media analytics framework used by video-centric applications.
Python code-based, containerized microservice for video analytics pipelines built using DL Streamer.
Tools#
Intel® Edge System Qualification
Benchmarks and qualifies Intel® hardware for optimal system selection, e.g., Smart Parking system.
Builds production-ready vision models for deployment in applications, e.g. Smart Parking.
Visual Pipeline And Platform Evaluation Tool
Benchmarks end-to-end visual analytics pipelines, e.g. Smart Parking pipelines, on Intel® hardware.
Fuses multi camera and sensor data into a spatial scene model, e.g. for Smart Parking scene.
Blueprints#
A blueprint for creating end-to-end pipelines powered by distributed AI agents for monitoring critical infrastructure. It includes implementations of a Vision AI and Multimodal use cases, both covering the common three-unit architecture, multi-agent reasoning layer, and deployment principles.
A blueprint for creating a digital twin of a building for operational efficiency, predictive
maintenance, and enhanced safety.
A demo will be available for download in the following weeks.