Mind the Domain Gap: Measuring the Domain Gap Between Real-World and Synthetic Point Clouds for Automated Driving Development

TL;DR

This paper introduces a new metric, DoGSS-PCL, to measure the domain gap between real and synthetic point clouds in automated driving, enabling better simulation credibility and training data generation.

Contribution

The paper proposes a novel metric for quantifying the domain gap between real and synthetic point clouds, facilitating comprehensive analysis and improved training for autonomous driving systems.

Findings

The proposed metric effectively measures the geometric and semantic differences.

Synthetic semantic point clouds can be used for training without significant performance loss.

The approach supports scalable and credible data simulation for automated driving.

Abstract

Owing to the typical long-tail data distribution issues, simulating domain-gap-free synthetic data is crucial in robotics, photogrammetry, and computer vision research. The fundamental challenge pertains to credibly measuring the difference between real and simulated data. Such a measure is vital for safety-critical applications, such as automated driving, where out-of-domain samples may impact a car's perception and cause fatal accidents. Previous work has commonly focused on simulating data on one scene and analyzing performance on a different, real-world scene, hampering the disjoint analysis of domain gap coming from networks' deficiencies, class definitions, and object representation. In this paper, we propose a novel approach to measuring the domain gap between the real world sensor observations and simulated data representing the same location, enabling comprehensive domain gap analysis. To measure such a domain gap, we introduce a novel metric DoGSS-PCL and evaluation assessing the geometric and semantic quality of the simulated point cloud. Our experiments corroborate that the introduced approach can be used to measure the domain gap. The tests also reveal that synthetic semantic point clouds may be used for training deep neural networks, maintaining the performance at the 50/50 real-to-synthetic ratio. We strongly believe that this work will facilitate research on credible data simulation and allow for at-scale deployment in automated driving testing and digital twinning.