Towards Safe Autonomous Driving: Capture Uncertainty in the Deep Neural Network For Lidar 3D Vehicle Detection

04/13/2018

∙

by   Di Feng, et al.

∙

0

∙

share

To assure that an autonomous car is driving safely on public roads, its deep learning-based object detector should not only predict correctly, but show its prediction confidence as well. In this work, we present practical methods to capture uncertainties in object detection for autonomous driving. We propose a probabilistic 3D vehicle detector for Lidar point clouds that can model both classification and spatial uncertainty. Experimental results show that our method captures reliable uncertainties related to the detection accuracy, vehicle distance and occlusion. The results also show that we can improve the detection performance by 1