Based on enormous data accumulation and real road analysis, our self-developed visual sensing system is able to precisely detect freespace and various traffic objects, including vehicles, pedestrians, lanes, traffic lights and sighs.
By combining neural-network -based visual sensing and sensor fusion, our in-cabin sensing technology is able to analyze the conditions and behaviors of passengers and drivers through monitoring eyelid, gazing direction, movement of head and body, and in-cabin objects.
Taking advantage of features of different sensors, our sensor fusion technology provides better sensing results for autonomous vehicles by target tracking, association and prediction.
Our neural network engine kit,FastNet, ThinNet and HardNet, is able to operate powerful deep learning model on low-cost and low-dissipation chips.