Abstract

How GPU Compute has Enabled Next Level Functionality and Robustness for Driver Monitoring

As autonomous vehicles make progress, it is important for cameras to better "understand" the driver, occupants, and the entire cabin. With the progress of deep learning, it is now possible to provide extremely accurate driver and cabin monitoring, in real time, using inexpensive camera sensors and available GPU ....


Session No: SIL8107
Speaker: Ophir Herbst
Type:

Autonomous Vehicle Technologies

Date: Thursday - October 18, 2018 01:00 PM - 01:45 PM
Location: Hall B
Topic: Autonomous Vehicle Technologies
Industry: Automotive / Transportation

As autonomous vehicles make progress, it is important for cameras to better "understand" the driver, occupants, and the entire cabin. With the progress of deep learning, it is now possible to provide extremely accurate driver and cabin monitoring, in real time, using inexpensive camera sensors and available GPU compute resources in the vehicle, leading to cost effective deployment options. We will show CoDriver, our end to end driver and cabin sensing software, and some of the innovations made available using AI/Deep learning. We will also demonstrate how these technologies transform the way car OEMs think about cabin sensing.