Smart Camera Reference Implementation

This repository is based on the Microsoft Smart Store Reference Architecture

（日本語版 README はこちらをご参照ください）

Overview

What Is a Smart Camera?

• It's an edge device which is based on a deep-learning inference engine for analyzing the video streams taken by itself or from a security camera. In this implementation, we use AI accelerators for edge devices to achieve the high-speed & high-accuracy person detection, person tracking, face detection and age/gender recognition tasks.
• The inference results sent from the edge device will be saved into Azure Cosmos DB, calculated by a server and displayed on a dashboard as the results of customer flow and age/gender distribution.
• The inference modules running on the edge device are managed by Azure IoT Hub. You can easily replace the modules by some other modules for different purpose such as product recognition to make your own Smart Camara system.

Getting Started

• To setup your edge device, follow this instruction: Deploy Azure IoT Device
• To setup a dashboard server, follow this instruction: Deploy Dashboard Server

More Details

• For the whole architecture of Smart Camera system: see this page.
• An example use case: see this page.
• Processing sequence and data flow graphs: see this page.