The NAIL Event-based Vision Benchmark Suite
Embedded Files
To facilite the development of event-based/-enhanced high-speed autonomy, we deliver novel challenging real-world event-based vision benchmarks. Our tasks of interest consist of state estimation (VO/VIO/SLAM), object-level Time-to-Contact (TTC) estimation and 3D object detection. To this end, we collected data (event stream, frames, and IMU measurements, etc.) that are tailored for high-speed and low-latency robotic applications. Besides providing all raw data, we provide a benchmark for each task. For each of our benchmarks, we also provide specific evaluation metrics and corresponding evaluation website.
Open Challenge Benchmarks: EvSLAM, TTC Estimation, 3D Object Detection.
Citation
You may need to cite the following papers when using the dataset and benchmark in your research.
For the raw data and evaluation metrics in the EvSLAM benchmark, please cite:
@article{zhong2026event,
title={Event-based SLAM Benchmark for High-Speed Maneuvers},
author={Zhong, Sheng and Niu, Junkai and Gallego, Guillermo and Sun, Kaizhen and Yi, Yang and Miao, Zhiqiang and Hu, Dewen and Wang, Yaonan and Scaramuzza, Davide and Zhou, Yi},
journal={arXiv preprint arXiv:2604.24033},
year={2026}
}
For the raw data and evaluation metrics in the TTC or 3D Detection benchmark, please cite:
@article{li2026toward,
title={Toward Deep Representation Learning for Event-Enhanced Visual Autonomous Perception: The eAP Dataset},
author={Li, Jinghang and Li, Shichao and Lian, Qing and Li, Peiliang and Chen, Xiaozhi and Zhou, Yi},
journal={IEEE Transactions on Robotics},
volume={42},
pages={1643--1661},
year={2026},
publisher={IEEE}
}
License
All datasets and benchmarks on this page are copyright by us and published under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License. This means that you must attribute the work in the manner specified by the authors, you may not use this work for commercial purposes and if you alter, transform, or build upon this work, you may distribute the resulting work only under the same license.
Page updated
Google Sites
Report abuse