R3ST (Realistic 3D Synthetic Trajectories)
收藏R3ST: A Synthetic 3D Dataset with Realistic Trajectories 数据集概述
基本信息
- 数据集名称:R3ST (Realistic 3D Synthetic Trajectories)
- 发布来源:Sapienza University of Rome, INSAIT (Sofia University), EURECOM
- 相关论文:R3ST: A Synthetic 3D Dataset with Realistic Trajectories
- 论文出处:Proceedings of Computer Analysis of Images and Patterns. CAIP 2025. Lecture Notes in Computer Science, vol 15622. Springer, Cham
- 发布年份:2026
数据集简介
R3ST是一个合成3D数据集,旨在通过生成合成3D环境并集成从真实无人机镜头记录的鸟瞰数据集SinD中提取的真实世界轨迹,来弥补合成数据与真实轨迹之间的差距。该数据集推进了道路车辆轨迹预测的研究,提供了准确的多模态真实标注和真实的人类驾驶车辆轨迹。
核心特点
- 真实轨迹:与通常由AI驱动或基于规则的算法决定车辆运动的典型合成数据集不同,R3ST集成了源自SinD数据集(一个具有从真实无人机镜头提取的精确车辆位置标注的鸟瞰数据集)中两个场景的真实世界车辆轨迹。
- 合成环境:通过使用Blender渲染创建的虚拟交叉口生成。
- 多模态标注:利用Vision Blender计算额外的多模态标注,可用于一系列计算机视觉应用。同时,直接从Blender世界环境中导出场景中每个对象的3D边界框,并将其投影到图像平面上以获得2D边界框。
应用与评估
- 主要应用:用于训练和评估用于交通分析及提高道路安全的计算机视觉模型,特别是在轨迹预测领域。
- 标注类型:提供实例分割和单目深度估计的定性结果示例。实例分割结果使用YOLO-Seg和SAM2在线演示获得,单目深度估计使用在KITTI上预训练的AnyDepth和Pixelformer Large执行。
获取与引用
-
数据集地址:https://r3st-website.vercel.app/
-
BibTeX引用:
@InProceedings{10.1007/978-3-032-05060-1_30, author="Teglia, Simone and Melis Tonti, Claudia and Pro, Francesco and Russo, Leonardo and Alfarano, Andrea and Pentassuglia, Matteo and Amerini, Irene", editor="Castrill{o}n-Santana, Modesto and Travieso-Gonz{a}lez, Carlos M. and Deniz Suarez, Oscar and Freire-Obreg{o}n, David and Hern{a}ndez-Sosa, Daniel and Lorenzo-Navarro, Javier and Santana, Oliverio J.", title="R3ST: A Synthetic 3D Dataset with Realistic Trajectories", booktitle="Computer Analysis of Images and Patterns", year="2026", publisher="Springer Nature Switzerland", address="Cham", pages="351--360", abstract="Datasets are essential to train and evaluate computer vision models used for traffic analysis and to enhance road safety. Existing real datasets fit real-world scenarios, capturing authentic road object behaviors, however, they typically lack precise ground-truth annotations. In contrast, synthetic datasets play a crucial role, allowing for the annotation of a large number of frames without additional costs or extra time. However, a general drawback of synthetic datasets is the lack of realistic vehicle motion, since trajectories are generated using AI models or rule-based systems. In this work, we introduce R3ST (Realistic 3D Synthetic Trajectories), a synthetic dataset that overcomes this limitation by generating a synthetic 3D environment and integrating real-world trajectories derived from SinD, a birds-eye-view dataset recorded from drone footage. The proposed dataset closes the gap between synthetic data and realistic trajectories, advancing the research in trajectory forecasting of road vehicles, offering both accurate multimodal ground-truth annotations and authentic human-driven vehicle trajectories. We publicly release our dataset here (https://r3st-website.vercel.app/).", isbn="978-3-032-05060-1" }




