The OU-ISIR Biometric Score Database

Introduction

The OU-ISIR Biometric Score Database is meant to aid research efforts in the general area of developing, testing and evaluating algorithms for biometric score-level fusion. The Institute of Scientific and Industrial Research (ISIR), Osaka University (OU) has copyright in the collection of scores and associated data and serves as a distributor of the OU-ISIR Biometric Score Database.

Format

Scores are provided in csv format of a dissimilarity score matrix, whose rows and cols correspond to probes and galleries, respectively. In addition, probe and gallery ID lists are also provided.

In the following sample, probe IDs in the Score.csv are 1, 1, 2, 2, 3, 3, 4, 4, 5, and 5 from top to bottom, while gallery IDs are 1, 2, 3, 4, and 5 from left to right. Hence, underlined scores are true match scores (so-called genuine or client scores), and the other scores are false match scores (so-called imposter scores).

Data collection

The data have been collected since March 2009 for video-based and sensor-based gait analysis. The approved informed consent was obtained from all the subjects in this dataset. Detailed descriptions of the video-based and sensor-based gait data are found in the following papers [1] and [2], respectively.

Set 1 (BSS1)

BSS1 is composed of dissimilarity score matrices from 1,936 probe subjects and 2,015 gallery subjects, which is a subset of OULP-V1C1-B75 defined in [1]. The individual dissimilarity scores are computed from the following gait features, which are extracted from the silhouette sequences of each subject. Detailed procedures of preprocess and feature extraction are found in [1].

Set 2 (BSS2)

BSS2 is composed of dissimilarity score matrices from 3,706 subjects, which is a full set of OULP-V1C1-A55 defined in [1]. The individual dissimilarity scores are computed from the following gait features, which are extracted from the silhouette sequences of each subject. Detailed procedures of preprocess and feature extraction are found in [1]

Set 3 (BSS3)

BSS3 is composed of dissimilarity score matrices from 737 subjects, which is a full set of the inertial sensor-based gait database defined in [2]. The individual dissimilarity scores are computed from the gait signals obtained by the following inertial sensors, which are attached to the each subject. Detailed procedures of preprocess and feature extraction are found in [2]

Set 4 (BSS4)

BSS4 is composed of dissimilarity score matrices from 3,249, which is a full set of OULP-V1C1-A85 and defined in [1]. The individual dissimilarity scores are computed from the following gait features, which are extracted from the silhouette sequences of each subject. Detailed procedures of preprocess and feature extraction are found in [1]

Set 5 (BSS5): MultiQ

BSS5 (MultiQ) is a single sensor-based multi-quality multi-modal biometric score database composed of dissimilarity score matrices of gait, head, and the height biometrics from 1,912 subjects as well as spatial and temporal resolutions as quality measures. This dataset contains a total of 130 combinations of 13 spatial resolutions and 10 temporal resolutions. Moreover, this dataset specifies training and test score matrices for cross validation. As a result, the database is quite large-scale (+3 billions scores!). Detailed procedures of preprocess and feature extraction as well as performance evaluation results are found in [3].

Set 6 (BSS6): MultiQ Score Database V2

BSS6 (MultiQ Score Database V2) is a very large-scale single sensor-based multi-quality multi-modal biometric score database to advance the research into evaluation, comparison, and benchmarking of score-level fusion approaches using both quality-independent and quality-dependent protocols. The database was constructed by using gait, head, and height modalities from the OU-ISIR Gait Database [1] and introduce spatial resolution (SR), temporal resolution (TR) and view as quality measures that significantly affect biometric system performance. It was considered seven and 10 scaling factors for SR and TR, respectively, with four view variations, as a result a database comprising approximately four million genuine and 7.5 billion imposter score databases was constructed. Detailed procedures of preprocess and feature extraction as well as performance evaluation results are found in [4].

How to get the database?

To advance the state-of-the-art in biometric score-level fusion, this database 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.

Note

It was often reported that Windows built-in zip did not work to extract all the files. Please try another unzip software if necessary.


The database administrator

Department of Intelligent Media, The Institute of Scientific and Industrial Research, Osaka University
Address: 8-1 Mihogaoka, Ibaraki, Osaka, 567-0047, JAPAN
Mail address
FAX: +81-6-6877-4375.