Embracing Uncertainties in 3D Object Detection -- from Annotation and Benchmarking to Failure Analysis and Data Augmentation
ABOUT THIS PROJECT
At a glance
|Masayoshi Tomizuka||Wei Zhan|
uncertainty, 3D detection, benchmark, data augmentation
Detecting objects with confidence is indispensable for safe autonomous driving. Though several methods have recently been proposed for uncertainty estimation in object detection using deep learning, there are still three important challenges towards the ideal probabilistic object detection.