The company, which offers deep learning-based vision processing for AVs, says that its SVNet perception software will be used in the demonstration vehicle, which uses Ambarella processors and software to enable cameras to become more intelligent by extracting data from high-resolution video streams. A fully-optimized SVNet will be ported on Ambarella's CV22 system-on-chip (SoC), and the demo will use both the SVNet External and SVNet Internal systems, supporting ADAS functions externally and the DMS internally.
This, says the company, is an example of the flexible nature of its software, as well as Ambarella's ability to accommodate sophisticated deep learning-based algorithms to execute complex functions necessary to operate autonomous vehicles (AVs).
"This demonstration will be simultaneously running both ADAS and DMS on one platform, with one network, a significant achievement in the industry," says StradVision CEO Juhnwan Kim. "Simultaneous ADAS and DMS will enable the industry to massively deploy advanced functions towards drivers' safety and improved driving experience, and we are excited to work with a respected brand like Ambarella to continue on this innovative path."
This, says the company, will enable the industry to deliver safe and reliable ADAS, AV and DMS perception solutions to the market in a realistic way, at a fraction of the cost - deploying not just computer vision, but deep learning-based vision that works in extreme conditions such as harsh weather and urban environments.
Ambarella CEO Fermi Wang adds, "We are pleased to be partnering with StradVision, a recognized leader in the field of AV vision processing software, for our automotive display during CES. Our collaboration at CES and beyond will lead to faster development and deployment of the next generation of AI-based ADAS systems."
Smart camera collaboration to develop next-gen ADAS
Automotive image sensors claim optimal autonomous driving performance
Automotive camera SoC combines AI with low power consumption
Ambarella, Smart Eye partner on next-gen AI-based driver monitoring