The Horned Sungem is a plug-and-play AI vision kit powered by the Intel Movidius MA245X convolutional neural network (CNN) accelerator. Connect via USB-C, run a brief installation script, and start performing real-time object detection, face recognition, and image analysis — no deep learning framework expertise required.

Compatible with TensorFlow, Caffe, Darknet, and other neural network frameworks, it delivers high-performance AI inference at under 3 W. Supports macOS, Linux (Ubuntu, CentOS, Fedora), Raspbian, Android, and Windows.

Key Features

• Intel Movidius MA245X CNN Accelerator – Hardware-accelerated neural network inference
• Plug and AI – USB-C connection with minimal setup for instant AI applications
• Built-In Camera – 1920×1080 or 640×360 resolution for real-time vision processing
• Multi-Framework Support – TensorFlow, Caffe, Darknet, and more
• Cross-Platform – macOS, Ubuntu, CentOS, Fedora, Raspbian, Android, Windows
• Ultra-Low Power – Less than 3 W consumption
• Lightweight – Under 10 g for easy integration into projects

Specifications

• Processor – Intel Movidius MA245X CNN accelerator
• RAM – 512 MB
• Interface – USB Type-C (USB 3.0)
• Camera – 1920×1080 or 640×360
• Power – Less than 3 W
• Weight – Less than 10 g
• Dimensions – 85.5 × 34 × 20 mm

Ideal For

• Learning and prototyping AI vision applications
• Object detection and face recognition projects
• Smart camera and intelligent robot development
• Adding AI inference to Raspberry Pi and other SBCs

Package Contents

• 1× Horned Sungem AI Vision Kit V1.0

Resources

• Quick Start Guide