Depression risk recognition based on gait: A benchmark
(1) Large Data Scale
The D-Gait dataset comprises 27,120 gait sequences collected from 292 volunteers. This extensive sample size provides robust support for data-driven methodologies.
(2) Comprehensive Diversity
The D-Gait dataset accounts for a wide range of shooting angles and clothing variations. Specifically, it features 16 uniformly distributed shooting angles from 0 to 180 degrees. Clothing variations include not only normal walking patterns but also factors such as walking with bags and walking in different clothing.
(3) Reliability of Label
The reliability of the D-Gait dataset labels is ensured through the integration of three professional diagnostic scales: SDS, PHQ-9, and GAD-7. This approach significantly enhances the accuracy and reliability of the labels.
This dataset includes silhouette and skeleton data versions, but it is essential to note that it is ACADEMIC USE ONLY.
To obtain and use this dataset and its subsets, all users are required to complete the following steps:
- Download the latest agreement and complete it.
- Submit it to BNU-IVC_D-Gait@outlook.com .
We will handle your requests within a week. In case you encounter any issues, please feel free to reach out to us via BNU-IVC_D-Gait@outlook.com.
Please cite the following paper if you find this useful in your research:
@article{liu2024depression,
title={Depression risk recognition based on gait: A benchmark},
author={Liu, Xiaotong and Li, Qiong and Hou, Saihui and Ren, Min and Hu, Xuecai and Huang, Yongzhen},
journal={Neurocomputing},
pages={128045},
year={2024},
publisher={Elsevier}
}