KY IR Shadow Gait Database

Introduction

Kyushu University Infrared Shadow Gait Database is composed of time-series shadow images for 54 subjects. All people walked straight, as indicated by dashed line 1 in the figure. To collect a shadow database, we used two infrared lights (Bosch, IR Illuminator 850 nm, UFLED30-8BD) and a camera (PointGrey Research Inc., Grasshopper2 M/C). The infrared lights were placed obliquely and the camera was placed on the ceiling perpendicular to the ground as shown in Fig. 1. Figure 2 an example movie.

Fig. 1 (a) Experimental setting, (b) actual scene.

Fig. 2 An example movei of IR shadow images.

Dataset

The KY IR Shadow Gait Database consists of time-series shadow silhouette images for 54 people. Each person has 6 sequences consiting of 3 different types of appearance:
- Type 1: Standard (4 sequences, Fig. 3)
- Type 2: Bag (1 sequence, Fig. 4 (a)-(c))
- Type 3: Cloth (1 sequence, Fig. 4 (d)-(f))
Figure 3 shows an example captured image and silhouette image. For more details, please refer to IR_shadow_appearance_types.xlsx.

Fig. 3 (a) Standard cloth, (b) an example of captured images.

Fig. 4 (a)-(c) Bag changes, (d)-(f) clothes changes.

This dataset was introduced in the EST 2014 paper, "Gait identification using invisible shadows: robustness to appearance changes" [1].

Download

You can download dataset from anonymous FTP server (ftp://robotics-ftp.ait.kyushu-u.ac.jp). All silhouette images are in "IR_Shadow" folder.

Citation

If you make use of the KY IR Shadow Gait Database in any form, please do cite the following paper [1]:

[1] Yumi Iwashita, Ryo Kurazume, Adrian Stoica, "Gait Identification Using Invisible Shadows: Robustness to Appearance Changes", Int. Conf. on Emerging Security Technologies, pp.34-39, September, 2014. doi: 10.1109/EST.2014.18

@inproceedings{yumi2014IRgait,
title={Gait Identification Using Invisible Shadows: Robustness to Appearance Changes},
author={Y. Iwashita and R. Kurazume and A. Stoica},
booktitle={Int. Conf. Emerging Security Technologies (EST)},
page={34-39},
year={2014},
month={September},
address={UK},
}

Updated 11/14/2016