The Software Innovation Award recognises a recent piece of software, technical team or company that has contributed most to the technical development of vehicle perception. Algolux was one of 6 finalists selected for the Award among a group of very competitive submissions.
Maximizing computer vision perception accuracy and robustness is critical for ADAS and autonomous vehicle safety, but good perception under everyday real-world driving conditions is quite difficult. Even today’s state of the art systems exhibit dramatically reduced accuracy when operated during dark nighttime conditions or in adverse weather. This means a vehicle, sign, or pedestrian may be missed, resulting in damage, injury, or even death, as we’ve tragically seen quite recently.
Algolux is a machine learning software company based in Montreal Canada that enables autonomous vision – empowering vehicles and devices to see more clearly and perceive what cannot be sensed with today’s vision systems. They’ve re-imagined today’s camera and perception approaches through CANA, a flexible end-to-end deep learning software architecture, to address this very problem and have demonstrated significantly improved accuracy against alternative perception systems. The software is quite scalable, able to deliver similar accuracy advantage whether on a big GPU or Nvidia Drive platform or down to a low-power sub 1-watt edge processor.
This allows these systems to not only improve safety across ADAS and autonomy, but to also be used in a much broader set of operation conditions than before to increase vehicle utilization, a key business value for autonomous “robo-taxi” fleets and industrial vehicles.
Another problem for the industry is a lack of consistent metrics to quantify image quality and perception accuracy across the difficult automotive use cases. This is now changing through the collaboration of leading automotive, computer vision, and imaging experts working within the IEEE P2020 Working Group. The organization is focused on establishing standard metrics and methods for automotive system image quality. Algolux has been a key contributor to the P2020 efforts, lending their deep expertise in imaging and perception to the effort.
Robert Stead, Awards Director for the AutoSens Awards shared his thoughts on Algolux being selected as a finalist, “Driving safety is paramount for ADAS and autonomous vehicles, and many companies are developing computer vision and deep learning techniques to improve perception. Algolux clearly stood out during the Award nomination process due to both the dramatic improvement in accuracy and robustness their groundbreaking neural network architecture provides and the flexibility and scalability of their approach. I can’t wait to see what happens on Award night.”
He continued, “Nominations for the Awards programme have almost doubled compared to last year’s event, so it was a difficult task to filter this down to the shortlist, but we are very happy to announce such an exciting line up of industry trail-blazers including those at the forefront of development of vehicle perception systems.”
The nominated finalists of each category will be scrutinised by our world-class judging panel over the summer, and the winners will be announced at The Atomium Brussels on the evening of 19 September 2018. There are a very limited number of tickets available for this intimate ceremony.
Algolux CEO Allan Benchetrit commented “We’re very honored to be recognized as a finalist for this year’s AutoSens Awards, after winning the “Most Exciting Startup” award at last year’s ceremony. It has been an incredible year for the company, with numerous customer engagements, recognition by a number of industry forums, and the closing of a Series A funding led by GM Ventures. These strongly validate the direction Algolux is going and the value our software brings to the industry.”
He further added “We see the collaborative and open dialog at AutoSens and the emphasis on standout innovation at the Awards as a way to accelerate progress on the difficult challenges of vehicle safety and enable new mobility models in the industry. It’s exciting to be part of it.”
