用少量真实数据+仿真数据，提升自动驾驶感知性能。
JiSAM: Alleviate Labeling Burden and Corner Case Problems in Autonomous Driving via Minimal Real-World Data
- 通过扰动增强与分域对齐，高效利用仿真数据
- 仅用2.5%真实标签即达全量真实数据性能
- 显著提升罕见物体检测，适合数据稀缺场景
基于深度学习的自动驾驶感知为安全环保交通带来希望。然而，依赖大量真实标注的激光雷达感知限制了实际道路测试规模。真实世界3D数据标注耗时耗能，且缺乏罕见交通参与者等极端情况。相比之下，CARLA等模拟器可轻松生成带标注的点云并包含极端案例。但将合成数据用于提升真实感知效果面临两大挑战：模拟数据样本效率低、仿真到真实的域差距。为此，我们提出一种即插即用方法JiSAM（Jittering augmentation, domain-aware backbone and memory-based Sectorized AlignMent）。在著名数据集NuScenes上的大量实验表明，使用SOTA 3D目标检测器，仅需2.5%的真实标签数据，即可实现与全量真实数据训练模型相当的性能；此外，对真实训练集中未标注的物体，仍取得超过15 mAP的提升。
原文摘要 · Abstract (English)
Deep-learning-based autonomous driving (AD) perception introduces a promising picture for safe and environment-friendly transportation. However, the over-reliance on real labeled data in LiDAR perception limits the scale of on-road attempts. 3D real world data is notoriously time-and-energy-consuming to annotate and lacks corner cases like rare traffic participants. On the contrary, in simulators like CARLA, generating labeled LiDAR point clouds with corner cases is a piece of cake. However, introducing synthetic point clouds to improve real perception is non-trivial. This stems from two challenges: 1) sample efficiency of simulation datasets 2) simulation-to-real gaps. To overcome both challenges, we propose a plug-and-play method called JiSAM , shorthand for Jittering augmentation, domain-aware backbone and memory-based Sectorized AlignMent. In extensive experiments conducted on the famous AD dataset NuScenes, we demonstrate that, with SOTA 3D object detector, JiSAM is able to utilize the simulation data and only labels on 2.5% available real data to achieve comparable performance to models trained on all real data. Additionally, JiSAM achieves more than 15 mAPs on the objects not labeled in the real training set.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。