The OU-ISIR Large Population Gait Database with real-life carried object and its performance evaluation

Abstract

In this paper, we describe the world's largest gait database with real-life carried objects (COs), which has been made publicly available for research purposes, and its application to the performance evaluation of vision-based gait recognition. Whereas existing databases for gait recognition include at most 4007 subjects, we constructed an extremely large-scale gait database that includes 62,528 subjects, with an equal distribution of males and females, and ages ranging from 2 to 95 years old. Moreover, whereas existing gait databases consider a few predefined CO positions on a subject's body, we constructed a database that contained unconstrained variations of COs being carried in unconstrained positions. Additionally, gait samples were manually classified into seven carrying status (CS) labels. The extremely large-scale gait database enabled us to evaluate recognition performance under cooperative and uncooperative settings, the impact of the training data size, the recognition difficulty level of the CS labels, and the possibility of the classification of CS labels. Particularly, the latter two performance evaluations have not been investigated in previous gait recognition studies.

Joint Metric Learning and Hierarchical Network for Gait Recognition

“…We use two datasets, CASIA-B [32] and OU-LP-Bag [33], to evaluate our method. CASIA-B and OU-LP-Bag are two public gait datasets.…”

Forensic Gait Analysis and Recognition: Standards of Evidence Admissibility

“…For example, a promising avenue is being taken by a team from Osaka University in Japan who have compiled the largest gait database of video and gait energy images (whole body shapes extracted from video) of 63,846 subjects (31,093 males and 32,753 females) with ages ranging from 2 to 90 years, the largest number of subjects being in the 6–15 years and 21–25 years age ranges and the lowest in the 51–90 years age range (approximately 2000 individuals in total) . Other large datasets also developed by this team include treadmill datasets which comprise speed and clothing variations and gait fluctuations , speed transition variation dataset , covariate (carrying objects) dataset , multiview camera angle dataset , and inertial sensors datasets . Although the datasets are varied and constitute a valuable step in gait recognition research, there are several limitations to consider.…”

“…In biometrics, gait is classified as a behavioral biometric , and the implication of this classification is that gait is not considered as individualistic as physiological biometrics (e.g., fingerprints), specifically because gait can be altered, albeit to a certain extent, by behavior. Inebriety, emotions, state of mind, and mood can all have an effect on a person's manner of walking, and implicitly, complicate the process of recognition, as do a variety of covariates such as clothing, shoes, carrying of items, type of walking surface, speed variation, among others . Thus, considering the large number of potential challenges which may arise, one may ask whether it is indeed worth considering gait as a biometric or even as a trait for forensic gait analysis.…”

View and Clothing Invariant Gait Recognition via 3D Human Semantic Folding

“…It is still a challenge to explore a GRU system involving a large population as most publicly available gait databases are limited to hundreds of subjects. However, it is worth noting that gait datasets involving large populations under different walking conditions have been published recently by Osaka University, i.e., OU-MVLP [6] with 14 views and 10,307 subjects, and OU-LP-Bag [7] with various carrying conditions and 62,528 subjects of all age ranges. As gait datasets involve larger populations, an emerging challenge is that the number of gait frames to be processed is typically enormous, requiring much processing time and storage space.…”