Edge AI Suite v1.0.0 (Initial Release)
Release Overview
The Edge AI Suite v1.0.0 marks our first public release, delivering collections of open, industry-specific AI software development kits (SDKs), microservices, and sample applications for independent software vendors (ISVs), system integrators, and solutions builders.
Key Goals
- The Metro AI Suite accelerates the development of solutions for Edge AI video, safety and security, smart city, and transportation use cases.
- The Manufacturing AI Suite enhances output quality and volume with automated defect detection, asset tracking, software-defined controls, and other capabilities, empowering smart factories with Vision, Time Series, and Generative AI capabilities.
- The Retail AI Suite accelerates hardware decisions for Retail AI workloads at the edge, featuring use cases such as self-checkout and loss prevention.
- The Media & Entertainment AI Suite provides libraries and sample applications to accelerate solution development for high-performance, high-quality, live video production, helping improve the viewer experience.
Key Components
The Edge AI Suites project hosts a collection of sample applications organized as follows:
| Suite | Sample Application | Get Started | Developers Docs |
|---|---|---|---|
| Metro AI Suite | Loitering Detection | Link | Customize the Application |
| Metro AI Suite | Image-Based Video Search | Link | Build and Deployment instructions |
| Metro AI Suite | Smart Parking | Link | Customize the Application |
| Manufacturing AI Suite | Pallet Defect Detection | Link | Customize the Application |
| Manufacturing AI Suite | Weld Porosity | Link | Customize the Application |
| Retail AI Suite | Automated Self Checkout | Link | Advanced Guide |
| Retail AI Suite | Loss Prevention | Link | Advanced Guide |
Highlighted Features
The Metro AI Suite includes the following tools and toolkits:
- Metro AI Suite SDK: Provides a comprehensive and modular toolkit for accelerated media processing and AI inference, designed to fast-track the development of visual AI solutions.
- Visual Pipeline and Performance Evaluation Tool: Assess Intel® hardware options, benchmark performance, and analyze key metrics to optimize hardware selection for AI workloads.
- Device qualification tool For device and equipment manufacturers, the suite includes a device qualification tool to verify and benchmark hardware performance and create system qualification reports. See the recommended hardware catalog for a list of qualified system configurations.
The Suite also provides a collection of visual analytics sample applications, using deep learning and large models (GenAI):
- Image Search by Text: A reference implementation using multi-modal large language models to perform image search with text query
- Visual Q&A: A reference implementation using multi-model large language models to perform Q&A on image data
- Loitering Detection: Effortlessly monitor and manage areas with AI-driven video analytics for real-time insights and enhanced security
- Image Based Video Search: Performs near real-time analysis and image-based search to detect and retrieve objects of interest in large video datasets.
- Smart Parking: Effortlessly manage parking spaces with AI-driven video analytics for real-time insights and enhanced efficiency.
The Manufacturing AI Suite helps you develop solutions for:
- Production Workflow: Efficiency optimizations, product quality (detect anomalies, defects, or variations)
- Workplace Safety: AI-based safety insights to help reduce risks
- Real-Time Insights: Improve the production process (local data processing, integration with existing manufacturing execution systems, tracking defect rates, identifying trends)
- Automation: Correct problems almost immediately (instant alerts, implementation of corrective actions)
The important components from Edge AI Libraries that help you develop such pipelines for industrial and manufacturing AI use cases are:
- Deep Learning Streamer Pipeline Server: Built on top of GStreamer, a containerized microservice for development and deployment of video analytics pipeline.
- Model Registry: Providing capabilities to manage lifecycle of an AI model.
- Time Series Analytics: Ability to build time-series based AI pipeline utilizing Kapacitor and Telegraf containers, with InfluxDB for data storage.
- Object Store Microservice: MinIO based object store microservice to build generative AI pipelines.
- Intel® Geti™ SDK: A python package containing tools to interact with an Intel® Geti™ server via the REST API, helping you build a full MLOps for vision based use cases.
The Retail AI Suite is an open-source software framework designed to accelerate AI workload evaluation and hardware selection for point-of-sale use cases at the edge. It helps retail solution builders assess device configurations across Intel® product generations to enhance decision-making and reduce the total cost of ownership. Key use cases include:
- Automated Self-Checkout: Product recognition (detection, classification, and tracking), full pipeline (product, weight, text, and barcode), age verification
- Loss Prevention: Fake scans, items in basket, multi-product identification, product switching, shopper behavior (obscuring/hiding an item), event video summation
Known Issues
None
Breaking Changes
None — this is the initial release.