SCUT FV Database
--__datasamples/original photo.
--__datasamples/ROI.
The SCUT FV Dataset contains 61,344 images from 568 fingers. For each finger, the first 6 insertions are in a normal posture, the other 12 insertions are in clockwise and counterclockwise rotations. The rotation angles of all images are <20°. For each insertions, 6 images are obtained under 6 NIR-light intensity levels illuminating the finger. Compared to other datasets, the images of the FV-SCUT dataset exhibit larger variation, making the training more challenging in terms of achieving good performance.
The SCUT FV Database is publicly available (free of charge) to the research community.
Unfortunately, due to privacy reasons, we cannot provide the database for commercial use.
We have made part of the dataset available for download in the repo in order to get a detailed view of this data. Those interested in obtaining the whole SCUT FV Database should download release agreement, and send by e-mail one signed and scanned copy to scutbip@outlook.com.
While reporting results using the SCUT FV Database, please cite the following article:
@article{Tang2019StudyOA,
title={Finger Vein Verification using a Siamese Convolutional Neural Network},
author={Su Tang and San Zhou and Wenxiong Kang and Qiuxia Wu and FeiQi Deng},
journal={IET Biometrics},
year={2019},
volume={8},
pages={306-315}
}
Prof. Kang Wenxiong
Biometrics and Intelligence Perception Lab.
College of Automation Science and Engineering
South China University of Technology
Wushan RD.,Tianhe District,Guangzhou,P.R.China,510641
auwxkang@scut.edu.cn