Eyeris launches in-vehicle scene understanding AI and interior image segmentation software
Eyeris Technologies, Inc. announced the commercial availability of its expanded product portfolio with an in-vehicle scene understanding (ISU) AI and interior image segmentation software designed to achieve the most accurate understanding of the automotive interior cabin space.
Eyeris' vision AI algorithms use multiple automotive-grade 2D RGB-IR image sensors to provide real-time analytics on the edge, powered by its proprietary AI chip. The expanded EyerisNet portfolio includes:
- Interior Image Segmentation, which provides a pixel map where every pixel in the vehicle interior scene is associated with a class label such as human, object, or surface, along with their corresponding regions and contours for greater interior scene understanding accuracy.
- Human Behavior Understanding (HBU) AI, which features body tracking, action, and activity recognition, face analytics, and emotion recognition for all occupants inside the vehicle.
- Object Localization, which provides detection, classification, size, and position of objects.
- Surface Classification, which enables identification and position of all in-cabin surfaces (such as footwells, door panels, center console, etc.), relative to occupants and objects.
"We continue to pioneer new technology capabilities to enhance safety, comfort, and convenience in autonomous and highly automated vehicles with our In-vehicle Scene Understanding AI," said Modar Alaoui, Founder and Chief Executive Officer of Eyeris. "The market lead we established earlier has further advanced significantly with the addition of Interior Image Segmentation technology, but more importantly, we are providing our customers with the capability to deliver a safer and more personalized experience for both privately-owned and shared vehicles."
Eyeris also introduced its proprietary AI chip—a compact, scalable hardware solution designed to inference the entire EyerisNet suite of deep neural networks from multiple 2D cameras, for real-time edge computing. The automotive-grade ASIC is AEC-Q100 qualified and consumes less than 7 watts.