The PKU-DVS dataset is constructed by National Engineering Laboratory for Video Technology (NELVT), Peking University, sponsored by the National Basic Research Program of China and Chinese National Natural Science Foundation. The goals to create the PKU-DVS include:
(1) providing the worldwide researchers of the spike coding community a large-scale DVS spike dataset for evaluating their coding algorithms;
(2) facilitating the development of high-efficient spike coding technologies by providing large-scale spike sequences with different depth of field or at different luminance and time conditions.
Therefore, the PKU-DVS dataset is now partly made available for the academic purpose only on a case-by-case basis. The first available part contains Class A (indoor) and Class B (outdoor) with seven and six spike sequences, respectively. More sequences are under evaluation which will be public available in the near future.

The PKU-DVS dataset is constructed by National Engineering Laboratory for Video Technology (NELVT), Peking University, sponsored by the National Basic Research Program of China and Chinese National Natural Science Foundation. The NELVT at Peking University is serving as the technical agent for distribution of the dataset and reserves the copyright of all the sequences in the dataset. Any researcher who requests the PKU-DVS dataset must sign this agreement and thereby agrees to observe the restrictions listed in this document. Failure to observe the restrictions will result in access being denied for the request of the future version of the PKU-DVS dataset and being subject to civil damages in the case of publication of sequences that have not been approved for release.

INTRODUCTION

The PKU-VD datasets including VD1 and VD2 are constructed by National Engineering Laboratory for Video Technology (NELVT), Peking University, sponsored by the National Basic Research Program of China and Chinese National Natural Science Foundation. We construct two large-scale vehicle datasets1 (i.e., VD1 and VD2) based on real-world unconstrained scenes from two cities respectively. The images in VD1 are obtained from high resolution traffic cameras, and images in VD2 are captured from surveillance videos. We perform vehicle detection on the raw data to make sure that each image only contains one vehicle. The region of plate number has been covered by black color due to privacy protection.

We provide diverse attribute annotations for each image in both two datasets, including identity number, precise vehicle model and vehicle color. Specifically, identity number (ID) is unique and all images belong to the same vehicle have the same ID (we make sure that there are at least two images in the dataset for each vehicle ID). We provide the most precise model type with detailed vehicle type and different produced years. For example, Audi-A6L-2012&2015, Audi-A6-2004, Audi-A4-2006&2008 and Audi-A4-2004&2005 are four different vehicle models in our datasets. As for color information, 11 common colors are annotated in our datasets. We carefully check all annotations to ensure the consistency of labels so that all the images belonging to the same vehicle ID are annotated with the same vehicle model and color. 

VD1: There are total 1,097,649 images in the dataset. We label 1,232 vehicle models and 11 colors.

VD2: There are total 807,260 images in the dataset. We label 1,112 vehicle models and 11 colors.
LICENSE 

The PKU-VD datasets are now partly made available for the academic purpose only on a case-by-case basis. The NELVT at Peking University is serving as the technical agent for distribution of the dataset and reserves the copyright of all the images in the dataset. Any researcher who requests the PKUVehicleID dataset must sign this agreement and thereby agrees to observe the restrictions listed in this document.

• The images and the corresponding annotation results for download are part of PKU-VD datasets.
• The images and the corresponding annotation results can only be used for ACADEMIC PURPOSES. NO COMERCIAL USE is allowed.
• Copyright © National Engineering Laboratory for Video Technology (NELVT) and Institute of Digital Media, Peking University (PKU-IDM). All rights reserved.

All publications using PKU-VD dataset should cite the paper below:

@inproceedings{yan2017exploiting,
 title={Exploiting Multi-Grain Ranking Constraints for Precisely Searching Visually-Similar Vehicles},
 author={Yan, Ke and Tian, Yonghong and Wang, Yaowei and Zeng, Wei and Huang, Tiejun},
 booktitle={Proceedings of the IEEE International Conference on Computer Vision},
 pages={562--570},
 year={2017}
}

DOWNLOAD

You can download the agreement (pdf) from here. Please make sure that a permanent/long-term responsible person (e.g., professor, PI) fills in the agreement with a handwriting signature. After filling it, please send the electrical version to our Email: pkuml at pku.edu.cn (Subject: PKU-VD-Agreement)

Please send it through an academic or institute email-addresses such as xxx at xxx.edu.xx. Requests from free email addresses (outlook, gmail, qq etc) will be kindly refused.

After confirming your information, we will send the download link and password to you via Email. You need to follow the agreement.

Usually we will reply in a week. But sometimes the mail does not arrive and display successfully for some unknown reason. If this happened, please change the content or title and try sending again.