Simulation-to-Reality domain adaptation for offline 3D object annotation on pointclouds with correlation alignment
This work addresses the annotation bottleneck for autonomous driving perception systems, offering an incremental improvement by adapting existing domain adaptation techniques to pointcloud data.
The paper tackles the costly problem of manually annotating 3D objects in LiDAR pointclouds for autonomous driving by proposing a semi-automatic method that uses simulated data from CARLA and real-world pointclouds, achieving domain adaptation through a CORAL loss to align feature representations.
Annotating objects with 3D bounding boxes in LiDAR pointclouds is a costly human driven process in an autonomous driving perception system. In this paper, we present a method to semi-automatically annotate real-world pointclouds collected by deployment vehicles using simulated data. We train a 3D object detector model on labeled simulated data from CARLA jointly with real world pointclouds from our target vehicle. The supervised object detection loss is augmented with a CORAL loss term to reduce the distance between labeled simulated and unlabeled real pointcloud feature representations. The goal here is to learn representations that are invariant to simulated (labeled) and real-world (unlabeled) target domains. We also provide an updated survey on domain adaptation methods for pointclouds.