Camera chip uses AI for monitoring applications

Camera chip uses AI for monitoring applications

New Products |
By Christoph Hammerschmidt

Ambarella, developer of video processing and computer vision semiconductors, introduces the camera System-on-Chip (SoC) CV25. This latest member of the CVflow family combines advanced image processing with powerful video coding and CVflow image processing in a single power-saving design. The CV25’s CVflow architecture provides the Deep Neural Network (DNN) processing required for the next generation of affordable and intelligent home surveillance, professional surveillance and automotive retrofit solutions, including intelligent dash cameras, driver monitoring systems (DMS) and electronic mirrors.

With the CV25, Ambarella aims to bring sophisticated computer vision to the mainstream. With the SoC, the company is focusing on reducing customers’ total system costs for providing significant image processing performance, high quality image processing and advanced cyber security features with very low power consumption. CV25-based cameras can apply artificial intelligence (AI), allowing functions such as real-time view recognition to be performed on the camera rather than in the cloud.

In driver monitoring systems, the CV25’s AI applications can detect driver fatigue or the level of distraction by monitoring his eyes and facial expression. In smart home applications, cameras equipped with the SV25 can detect people and distinguish between pets, people and vehicles. In intelligent video doorbells, CV25 can automatically detect known faces approaching the front door, identify unknown people and notify the homeowner when a package is delivered.

The CV25 offers AVC and HEVC video encoding with very low bit rates to minimize cloud storage costs. A high-performance image signal processor (ISP) delivers outstanding image quality even in poor lighting conditions, promises Ambarella. High Dynamic Range (HDR) processing ensures maximum image detail in high-contrast scenes. It includes a comprehensive suite of advanced cyber security features to protect against hacking, including Secure Boot, TrustZone and I/O virtualization. Based on 10nm ultra-low power process technology, the CV25 chip is optimized for wireless camera applications that require long battery life and small form factors.

As part of the CVflow family, the CV25 chip features a common SDK, Computer Vision (CV) tools, ISP and cyber security features such as the existing CV22 and CV2 SoCs, enabling multiple price/performance options. With Ambarella’s complete CV tools, customers can easily port their own neural networks to CV25. The set includes a compiler, debugger and support for industry standard machine learning frameworks such as Caffe and TensorFlow with comprehensive guidelines for optimizing the performance of Convolutional Neural Networks (CNN).

More information:


If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles