Venus AI Box

Our Mission

Enabled by Ambrella CV2 advanced computer chip

Our Venus AI Box is powered by the latest and greatest Ambrella CV2 advanced computer chip which packs an incredible punch when it comes to raw processing power. Successfully, leveraging Ambrella CV2’s complete SoC package, Leopard Imaging managed to come up with a highly functional, super cost-efficient solution to your computer vision and image processing needs with Venus AI Box.

Supports all deep learning and visual AI applications

Venus AI Box was designed with excellent processing performance and high capability in mind. Thanks to Ambrella CV2 being 20 times more powerful than its previous generation CV1, this single device can support advanced computer vision, stereovision, 60fps video encoding at 4k resolution, and all the image processing you can imagine. Basically, it can handle virtually anything you throw at it!

Outstanding low-light performance with HDR processing

Ambrella CV2’s incredibly powerful architecture provides a robust image processing pipeline for the best results. Leopard Imaging’s Venus AI Box makes good use of this as our advanced HDR (High Dynamic Range) processing technology offers brilliant low-light performance and delivers high-quality imaging even in challenging lighting conditions with high contrast. Venus AI Box is the go-to option if your deep-learning data requires better-quality images.

Revolutionizing automotive ADAS and self-driving systems

Venus AI Box supports highly efficient 60fps AVC and HEVC video encoding at 4k resolution. This enables the addition of high-quality video data in automotive ADAS and self-driving systems which paves the way for the development of multi stream and multi-imager IP security cameras. With this powerful device, self-driving systems will be more able and safer than ever, thanks to high quality, readily available imaging data due to higher fps.

Supports all deep learning and visual AI applications

Ambrella CV2 offers best-in-class on-chip security features that prevent your system from being hacked by malicious agents. It includes a full suite of advanced security features such as secure boot, TrustZone™, and I/O virtualization.

Port your own neural networks in Venus AI Box

Leopard Imaging provides an elaborate and complete set of tools to enable customers to easily port their custom neural networks onto the CV2 SoC. This set of tools includes a compiler, a debugger, and professional support for industry-standard training tools including Caffe™ and TensorFlow™. Our support also provides extensive and in-depth guidelines for CNN (Convolutional Neural Network) performance optimizations for your application.

Applications

ADAS
Robotics
Drones
IoT

Key Features

Multi-channel ISP with up to 800
Mpixel/s input pixel rate
Deep Learning Support
1.2 GHz CPU with
1MB L2 ca
16 GB Internal NAND
Flash 1 Micro SD Card interface
Memory: 4-GByte LPDDR4/4X
SDRAM @ 1.8 GHZZ

Questions?

Contact Us