Decision-level fusion for single-view gait recognition with various carrying and clothing conditions

Al-Tayyan, Amer; Assaleh, Khaled; Shanableh, Tamer

doi:10.1016/j.imavis.2017.02.004

Cited by 29 publications

(14 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, view-invariance to achieve more reliable gait recognition has been studied by several research groups [19,33,34,43,58,69]. Clothing and carrying conditions are other important covariate factors that are frequently investigated [2,24,53].…”

Section: Background and Relevant Workmentioning

confidence: 99%

“…A is equal to the outer product of N vectors A = u (1) • u (2) (N) , which means that for all values of indexes, A(i 1 , i 2 , . .…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

“…. , L, andŨ (1) ∈ R I 1 ×P 1 is a projection matrix along the first mode of the tensor, and similarly forŨ (2) andŨ (3) , such that the total tensor scatter…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

“…n column ofŨ (n) . In the Q-based method [45], for each n, the first P n eigenvectors are kept in the n-mode such that Q (1) = Q (2) =, . .…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

See 3 more Smart Citations

Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition

Kwolek

Michalczuk

Krzeszowski

et al. 2019

Multimed Tools Appl

View full text Add to dashboard Cite

We introduce synchronized and calibrated multi-view video and motion capture dataset for motion analysis and gait identification. The 3D gait dataset consists of 166 data sequences with 32 people. In 128 data sequences, each of 32 individuals was dressed in his/her clothes, in 24 data sequences, 6 of 32 performers changed clothes, and in 14 data sequences, 7 of the performers had a backpack on his/her back. In a single recording session, every performer walked from right to left, then from left to right, and afterwards on the diagonal from upperright to bottom-left and from bottom-left to upper-right corner of a rectangular scene. We demonstrate that a baseline algorithm achieves promising results in a challenging scenario, in which gallery/training data were collected in walks perpendicular/facing to the cameras, whereas the probe/testing data were collected in diagonal walks. We compare performances of biometric gait recognition that were achieved on marker-less and marker-based 3D data. We present recognition performances, which were achieved by a convolutional neural network and classic classifiers operating on gait signatures obtained by multilinear principal component analysis. The availability of synchronized multi-view image sequences with 3D locations of body markers creates a number of possibilities for extraction of discriminative gait signatures. The gait data are available at http://bytom.pja.edu.pl/projekty/hm-gpjatk/.

show abstract

Section: Background and Relevant Workmentioning

confidence: 99%

“…A is equal to the outer product of N vectors A = u (1) • u (2) (N) , which means that for all values of indexes, A(i 1 , i 2 , . .…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

“…. , L, andŨ (1) ∈ R I 1 ×P 1 is a projection matrix along the first mode of the tensor, and similarly forŨ (2) andŨ (3) , such that the total tensor scatter…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

“…n column ofŨ (n) . In the Q-based method [45], for each n, the first P n eigenvectors are kept in the n-mode such that Q (1) = Q (2) =, . .…”

Section: I N ) • U(j N I N ) a Rank-1 Tensormentioning

confidence: 99%

See 2 more Smart Citations

Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition

Kwolek

Michalczuk

Krzeszowski

et al. 2019

Multimed Tools Appl

View full text Add to dashboard Cite

show abstract

“…For instance, while completing a project on human-robot interaction several years ago, an undergraduate student at the Massachusetts Institute of Technology (MIT) discovered racial bias of a generic facial recognition algorithm which failed to recognize individuals of African American ancestry; when interacting with the robot which was equipped with a camera and a facial recognition software, the incorporated algorithm failed to recognize the face of the student, and implicitly, her presence (112). A similar issue is present (though yet to be acknowledged research-wise) in gait recognition studies, where gait data are predominantly obtained from a small number of participants (notable exceptions are the Osaka University datasets), usually adult males of 20-35 years of age with average height and weight, with ancestral background not being reported (e.g., [113][114][115]). However, even with large datasets, no extrapolations can yet be made regarding how and to what extent are these algorithms designed to include and process human variability in different populations, how their design influences the decision-making process of the biometric system and to what extent other human cognitive biases are inadvertently introduced into their design.…”

Section: The Position Of Gait Analysis and Recognition In The Forensimentioning

confidence: 99%

Forensic Gait Analysis and Recognition: Standards of Evidence Admissibility

Macoveciuc

Rando

Borrion

2019

Journal of Forensic Sciences

View full text Add to dashboard Cite

Gait is one biological characteristic which has attracted strong research interest due to its potential use in human identification. Although almost two decades have passed since a forensic gait expert has testified to the identity of a perpetrator in court, the methods remain insufficiently robust, considering the recent paradigm shift witnessed in the forensic science community regarding quality of evidence. In contrast, technological advancements have taken the lead, and research into automated gait recognition has greatly surpassed forensic gait analysis in terms of the size of acquired datasets and demographic variability of participants, tested variables, and statistical evaluation of results. Despite these advantages, gait recognition presents with different problems which are yet to be resolved. Therefore, courts should treat gait evidence with caution, as they should any other form of evidence originating from disciplines without fully established codes of practice, error rates, and demonstrable applications in forensic scenarios.

show abstract

On Human Identification Using Running Patterns: A Straightforward Approach

Anusha

Jaidhar

2019

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

Decision-level fusion for single-view gait recognition with various carrying and clothing conditions

Cited by 29 publications

References 49 publications

Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition

Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition

Forensic Gait Analysis and Recognition: Standards of Evidence Admissibility

On Human Identification Using Running Patterns: A Straightforward Approach

Contact Info

Product

Resources

About