Algolux Brings Atlas to the Cloud to Democratize Camera ISP Optimization for Computer Vision

Atlas optimizes cameras at scale for vision system design teams while maximizing computer vision accuracy and image quality.
However, camera development currently relies on expert imaging teams or external image quality service companies to manually tune camera architectures.
As such, this process does not ensure that the camera provides the optimal output for image quality or computer vision applications.
Atlas in the cloud and expanded ISP ecosystem support enables project scalability, optimal image quality and computer vision results, and lower program cost and schedule risks," said Allan Benchetrit, CEO of Algolux.

MONTREAL, April 6, 2021 /PRNewswire-PRWeb/ -- Algolux, the leading provider of robust and scalable perception solutions, has announced the next generation of its Atlas Camera Optimization Suite, now enabled in the cloud and supporting an extended set of camera image signal processors (ISPs) from Arm and Renesas.

Atlas optimizes cameras at scale for vision system design teams while maximizing computer vision accuracy and image quality. Flexible and secure cloud computing enables a more effective use of engineering resources to achieve highly optimized results for existing vision systems or those in-development. Additionally, SoC / ISPs providers can automate and significantly scale their ability to support customer vision system programs, shrinking customers' time to market and achieving optimal system performance. By automating ISP tuning through machine learning, the process is reduced from months to just a few days.

Optimizing safety-critical vision systems securely at scale

Cameras are the sensor of choice for system developers of safety-critical applications, such as automotive ADAS, autonomous vehicles and robots, or video security. However, camera development currently relies on expert imaging teams or external image quality service companies to manually tune camera architectures.

This painstaking approach can take months, requires hard-to-find deep expertise, and is visually subjective. As such, this process does not ensure that the camera provides the optimal output for image quality or computer vision applications.

The new cloud-enabled Atlas automated workflow further shortens today's manual ISP tuning process by up to 100x while providing the flexibility to adapt to requirement changes and more securely protect their confidential vision system details and algorithmic IP. The workflow allows teams to better scale their resources across programs, quickly evaluate different lens/sensor options for either maximum performance or cost reduction, and determine optimal camera ISP parameters for computer vision tasks across specific use cases or operating conditions, something not possible with today's manual tuning process.

Results in a recently published case study with a leading automotive Tier 1 show Atlas improving their object detection accuracy by 3 to 48 mAP points vs. the manual image quality tuned baseline of their Sony IMX490 sensor and Renesas V3H ISP camera system.

For providers of vision system SoCs and ISPs, Atlas in the cloud can be made accessible to their vision system customers as a predictable always-available resource that automatically optimizes the ISP for their specific configurations and requirements. This dramatically improves effectiveness and scalability of support, shrinks turn-around time, and delivers consistent and optimal results.

Support for Arm and Renesas ISPs

Atlas now supports an expanded set of ISPs for automotive and IoT applications:

The Arm Mali-C71AE and Mali-C52 ISPs for smart automotive and IoT systems (see related release "Algolux Brings Atlas Camera ISP Optimization for Computer Vision to Arm ISP Users".)
ISPs embedded in Renesas R-Car SoCs, such as the R-Car V3H and R-Car V3M for intelligent and automated driving (AD) vehicles, and the recently announced V3U ASIL D SoC for advanced driver assistance systems (ADAS) and AD systems (see related release "Algolux Collaborates with Renesas on R-Car Camera Optimization for Computer Vision".)

Atlas support for additional ISPs will be announced later this year.

"Vision system teams struggle to predictably achieve good results with today's lengthy manual camera tuning processes, fixed deadlines, and limited in-house or outsourced resources. Atlas in the cloud and expanded ISP ecosystem support enables project scalability, optimal image quality and computer vision results, and lower program cost and schedule risks," said Allan Benchetrit, CEO of Algolux.

About Algolux

Algolux is an award-winning AI software company delivering the industry's most robust and scalable perception for all conditions, addressing both existing cameras and new designs through cloud-based tools and embedded software. The company was founded on groundbreaking research at the intersection of deep learning, computer vision, and computational imaging. Our computer vision and image optimization solutions address the mission-critical issue of safety for automotive ADAS, autonomous vehicles, fleets, autonomous mobile robots, and video security.

Algolux is headquartered in Montreal, with offices in Palo Alto and Munich, and has over 50 employees (85% in R&D). The company has numerous engagements spanning automotive, AVs, fleet management, and video security with leading customers worldwide.

Learn more at algolux.com

For the latest news on Algolux, follow us on LinkedIn, Twitter and YouTube.

Media Contact

Max Leblond, Algolux, 514-589-3933, [email protected]

Twitter

SOURCE Algolux