The data is split into training and test set. While ground truth data is available for the training data, it is retained for the test data with the exception of the semantic segmentation masks.
Dataset files are provided in larger files of multiple sequences or individual files per sequence.
Multi-Sequence Files
These files contain the whole dataset in larger files to simplify the download.
Training Data
train_events.zip (125 GB)
train_images.zip (216 GB)
train_disparity.zip (12 GB)
train_optical_flow.zip (3.7 GB)
train_semantic_segmentation.zip (88.6 MB)
train_calibration.zip
Test Data
test_events.zip (27 GB)
test_images.zip (43 GB)
test_semantic_segmentation.zip (28.9 MB)
test_calibration.zip
Lidar and Imu Data (Train + Test)
Contains Lidar and IMU data in rosbag format together with the calibration files.
Evaluation
The files contain timestamps and corresponding image indices that specify when a prediction is evaluated with the corresponding ground truth by the evaluation script.
test_disparity_timestamps.zip
test_foward_optical_flow_timestamps.zip
Citation
When using this work in an academic context, please cite the following publications:
@Article{Gehrig21ral,
author = {Mathias Gehrig and Willem Aarents and Daniel Gehrig and Davide Scaramuzza},
title = {DSEC: A Stereo Event Camera Dataset for Driving Scenarios},
journal = {IEEE Robotics and Automation Letters},
year = {2021},
doi = {10.1109/LRA.2021.3068942}
}@Article{Gehrig21threedv,
author = {Mathias Gehrig and Mario Millh\"ausler and Daniel Gehrig and
Davide Scaramuzza},
title = {{E}-{RAFT}: Dense Optical Flow from Event Cameras},
journal = {International Conference on 3D Vision (3DV)},
year = 2021
}When using the semantic segmentation masks in an academic context, please cite additionally the following publication:
@Article{Sun22eccv,
author = {Zhaoning Sun* and Nico Messikommer* and Daniel Gehrig and Davide Scaramuzza},
title = {ESS: Learning Event-based Semantic Segmentation from Still Images},
journal = {European Conference on Computer Vision. (ECCV)},
year = {2022},
}Individual Files
The following files contain the whole dataset but in individual files per sequence.
