OU-ISIR Gait Database, Multi-View Large Population Dataset with Pose Sequence

Introduction

The OU-ISIR Gait Database, Multi-View Large Population Database with Pose Sequence (OUMVLP-Pose) is meant to aid research efforts in the general area of developing, testing and evaluating algorithms for model-based gait recognition. The Institute of Scientific and Industrial Research (ISIR), Osaka University (OU) has copyright in the collection of gait video and associated data and serves as a distributor of the OUMVLP-Pose dataset.

This data set was built upon OU-MVLP. It contains 10,307 subjects of round-trip walking sequences captured by seven network cameras at intervals of 15° (this sums to 14 views by considering the round trip on the same walking course) with an image size of 1,280 x 980 pixels and a frame-rate of 25 fps. The video capturing setup is shown below.



The pose sequences are then extracted from RGB images using two state-of-the-art pose estimation algorithms, i.e., OpenPose [Cao+ CVPR2017] and AlphaPose [Fang+ ICCV2017]. The estimated pose results include 18 joints, as an example sequence at 45° shown below.



The number of frames in a sequence is from 18 to 35, and most of the sequences contain approximately 25 frames.

The detailed descriptions are found in the following paper.

  • W. An, S. Yu, Y. Makihara, X. Wu, C. Xu, Y. Yu, R. Liao, Y. Yagi, ``Performance Evaluation of Model-based Gait on Multi-view Very Large Population Database with Pose Sequences,'' IEEE Trans. on Biometrics, Behavior, and Identity Science, Vol. 2, No. 4, pp. 421-430, Oct. 2020 (Published online: 13 July 2020). [open access] [Bib]

Dataset

Two data sets are provided for OUMVLP-Pose, which were obtained by the OpenPose and AlphaPose, respectively. These two datasets contain the same number of subjects and the same parameters. For each of them, the entire data set was divided into two disjoint subsets, i.e., training and testing set, that both have almost the same size.

How to get the dataset?

To advance the state-of-the-art in gait-based application, this dataset including a set of pose sequences and subject ID lists of training and testing set could be downloaded as a zip file with password protection and the password will be issued on a case-by-case basis. To receive the password, the requestor must send the release agreement signed by a legal representative of your institution (e.g., your supervisor if you are a student) to the database administrator by mail, e-mail, or FAX.