arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.07310cs.CV
收藏

轨迹预测中跨数据集性能的不确定性估计

Uncertainty estimation for Cross-dataset performance in Trajectory prediction

Huawei Technologies France(华为技术法国公司);Mines ParisTech(巴黎矿业学院)

Thomas Gilles, Stefano Sabatini, Dzmitry Tsishkou, Bogdan Stanciulescu, Fabien Moutarde

摘要:

虽然已经进行了大量的工作,发展轨迹预测方法,并提出了各种数据集的基准测试这项任务,很少有研究已经做了这些方法的通用性和跨数据集的可移植性。在本文中,我们观察了两种最新的最先进的轨迹预测方法在四个不同数据集(Argoverse,NuScenes,Interaction,Shifts)上的性能。该分析允许获得关于最新轨迹预测模型的可推广性特性的一些见解,并分析哪个数据集更能代表真实驾驶场景,从而实现更好的可移植性。此外,我们提出了一种新的方法来估计预测不确定性,并展示了如何使用它来实现更好的跨数据集的性能。

英文摘要:

While a lot of work has been carried on developing trajectory prediction methods, and various datasets have been proposed for benchmarking this task, little study has been done so far on the generalizability and the transferability of these methods across dataset. In this paper, we observe the performance of two of the latest state-of-the-art trajectory prediction methods across four different datasets (Argoverse, NuScenes, Interaction, Shifts). This analysis allows to gain some insights on the generalizability proprieties of most recent trajectory prediction models and to analyze which dataset is more representative of real driving scenes and therefore enables better transferability. Furthermore we present a novel method to estimate prediction uncertainty and show how it could be used to achieve better performance across datasets.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑