Home >> Partnerships and Alliances >> Device Partnerships >> Ambarella
Ambarella and eInfochips –
Accelerating AI Innovation
PARTNERSHIP OVERVIEW
Founded in 2004, Ambarella, Inc. (NASDAQ: AMBA) is an edge AI semiconductor company whose products are used in a wide variety of human and computer vision applications, including automotive (ADAS/AD/DMS/OMS/eMirrors), robotics (AMRs, consumer robots, machine vision), security & surveillance, access control, and consumer & commercial. Ambarella’s low-power systems-on-chip (SoCs) offer advanced image processing, powerful deep neural network processing and high-resolution video compression to enable intelligent perception, fusion and central processing systems to extract valuable data from high-resolution video and radar streams.
eInfochips partnered with Ambarella to provide product development and engineering services for companies creating the next generation of AI perception products in the below key markets. Companies looking to accelerate their designs can leverage eInfochips’ deep experience to innovate and take full advantage of Ambarella SoCs, which feature:
FEATURED AI SoCs
CV2
Designed for accurate 3D environmental modeling and real-time neural network performance, the CV2 chip enables advanced robotics and industrial applications, including autonomous robots.
Show More
Key Features
- CNN- / DNN-based processing: detection, classification, tracking, and more
- Dedicated CVflow computer vision processor
- Tools for high- and low-level algorithm development
- Dedicated hardware for stereo processing
- CNN toolkit for easy porting with Caffe, TensorFlow, and ONNX
- Advanced image processing, with HDR, dewarping, EIS, MCTF, superior low-light imaging, and more
- High-efficiency video encoding up to 4KP60 performance
- Secure boot with TrustZone® and secure memory, true random number generator (TRNG), one-time programmable memory (OTP), DRAM scrambling and virtualization
CV5
Ambarella’s CV5 combines 8K image processing, video encoding / decoding, and CVflow® computer vision processing in a single, low-power design. Fabricated in advanced 5nm process technology, CV5 achieves power consumption below 4W for 8K video recording.
Show More
Key Features
- 8KP30 video encoding performance with multi-streaming capability
- CNN toolkit for easy porting of neural networks implemented in Caffe, TensorFlow, PyTorch, or ONNX frameworks
- Superior low-light processing
CV22
Ambarella’s CV22 SoC combines advanced image processing with our powerful CVflow® computer vision architecture to deliver the next generation of intelligent robotics and industrial applications.
Show More
Key Features
- CNN- / DNN-based processing: detection, classification, tracking, and more
- Dedicated CVflow computer vision processor
- Tools for high- and low-level algorithm development
- CNN toolkit for easy porting with Caffe, TensorFlow, and ONNX
- Advanced image processing with HDR, dewarping, EIS, MCTF, superior low-light imaging, and more
- High-efficiency video encoding up to 4KP30+ performance
- Secure boot with TrustZone® and secure memory, true random number generator (TRNG), one-time programmable memory (OTP), DRAM scrambling and virtualization
CV25
The CV25 chip combines advanced image processing and real-time CVflow® computer vision performance for intelligent industrial/robotics applications.
Show More
Key Features
- CNN- / DNN-based processing: detection, classification, tracking, and more
- Dedicated CVflow computer vision processor
- Tools for high- and low-level algorithm development
- CNN toolkit for easy porting with Caffe, TensorFlow, and ONNX
- Advanced image processing with HDR, dewarping, EIS, MCTF, superior low-light imaging, and more
- High-efficiency video encoding up to 4KP30+ performance
- Secure boot with TrustZone® and secure memory, true random number generator (TRNG), one-time programmable memory (OTP), DRAM scrambling and virtualization
CV28M
Our CV28M SoC is designed for intelligent applications at the edge, including smart home security, retail monitoring, consumer robotics, and occupancy monitoring. The chip’s high-performance CVflow® architecture provides deep neural network (DNN).
Show More
Key Features
- CVflow computer vision engine
- CNN / DNN inference acceleration for detection, classification, and more
- CNN toolkit for easy porting with Caffe, TensorFlow, and ONNX
- Up to 4KP30 video performance
- Advanced image processing with HDR, dewarping, EIS, MCTF, superior low-light imaging, and more
- Secure boot with TrustZone® and secure memory, true random number generator (TRNG), one-time programmable memory (OTP), DRAM scrambling and virtualization
EVALUATION KITS
CV2
“Chestnut” EVK
CV5
“Titanium” EVK
CV22
“Walnut” EVK
CV25
“Hazelnut” EVK
CV28
“Cashewnut” EVK
EXAMPLE Camera Markets & Applications
ROBOTICS
- Consumer Robots
- Industrial Robots
- AMR/Factory Automation
- Machine Vision
ACCESS CONTROL
- Door Locks
- Face ID Readers
SECURITY & SURVEILLANCE
- 2MP to 4K AI Cameras
- Single and Multi-Imagers
- Advanced AI-Based Analytics
CONSUMER & COMMERCIAL
- Smart Appliances
- Video Conferencing
- Smart Retail
eINFOCHIPS ENGINEERING SERVICES
This strategic partnership with Ambarella leverages eInfochips’ strong experience in camera design to provide product design and development services for a wide variety of camera solutions, including…
• End-to-end design and development based on Ambarella’s
AI SoC families
• Multi-sensor camera designs and video analytics
• Camera SDK and system software development
• Image sensor tuning in our state-of-the-art image tuning lab
• AI/ML algorithm porting and machine vision enablement
• Prototype and manufacturing support
WHY eINFOCHIPS?
- 500 plus products developed for our customers with 40 million deployments
- Experience across a wide variety of camera technologies and protocols, including: RGB IR cameras, PTZ, ONVIF, PSIA, ZigBee wireless, on-camera DVRs, multi-imager designs
- State-of-the-art Image tuning lab, and experience with the latest image sensors
- Platform-agnostic, reusable camera framework for faster development
- Proven IP, solution accelerators, multimedia codecs, protocol stacks, and an analytics library
