We develop AI based solutions primarily, but not exclusively targeted to, automotive applications. Our object detection algorithms, running on a power and cost efficient embedded platforms are suitable for automotive and industrial applications.
Algorithms based on Convolutional Neural Networks provide various types of collision alerts as well as detections of features required to implement situational awareness or ADAS systems. Our algorithms can detect: road signs and signals, cars, trucks, people, bicycles, road lanes, obstacles and others.
Our team of experts with extensive experience in camera image processing field continuously improve our algorithms, but also develop innovative and efficient testing and data collection tools and technologies.
Video below has been captured during tests of the ADAS system that we develop for one of our customers. It presents visualization of Forward Collision Warning feature (FCW)
Next video clip shows FCW warning together with line markings detection.
Following clip shows how LDW feature performs on the highway
Clip below shows driver monitoring models including: head pose estimation, distraction detection, eye gaze estimation, driver id and drowsiness detection.
