iLGaCo: Incremental Learning of Gait Covariate Factors

Mu, Zihao1; Castro, Francisco M.2; Marin-Jimenez, Manuel J.3; Guil, Nicolas2; Li, Yan-Ran1,4; Yu, Shiqi5

10.1109/IJCB48548.2020.9304857

2020

IEEE/IAPR International Joint Conference on Biometrics (IJCB)

2474-9680

978-1-7281-9187-4

IEEE/IAPR INTERNATIONAL JOINT CONFERENCE ON BIOMETRICS (IJCB 2020)

1-8

SEP 28-OCT 01, 2020

null,null,ELECTR NETWORK

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE

Gait is a popular biometric pattern used for identifying people based on their way of walking. Traditionally, gait recognition approaches based on deep learning are trained using the whole training dataset. In fact, if new data (classes, view-points, walking conditions, etc.) need to be included, it is necessary to re-train again the model with old and new data samples. In this paper, we propose iLGaCo, the first incremental learning approach of covariate factors for gait recognition, where the deep model can be updated with new information without re-training it from scratch by using the whole dataset. Instead, our approach performs a shorter training process with the new data and a small subset of previous samples. This way, our model learns new information while retaining previous knowledge. We evaluate iLGaCo on CASIA-B dataset in two incremental ways: adding new view-points and adding new walking conditions. In both cases, our results are close to the classical 'training-from-scratch' approach, obtaining a marginal drop in accuracy ranging from 0.2% to 1.2%, what shows the efficacy of our approach. In addition, the comparison of iLGaCo with other incremental learning methods, such as LwF and iCarl, shows a significant improvement in accuracy, between 6% and 15% depending on the experiment.

Training Gait recognition Adaptation models Feature extraction Legged locomotion Task analysis Biological system modeling

其他

英语

EI ; CPCI-S

Spanish Ministry of Science and Technology["TIN2016-80920R","RED2018102511-T"]

Computer Science ; Engineering ; Mathematical & Computational Biology

Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic ; Mathematical & Computational Biology

WOS:000723870900004

20210409828150

Biometrics ; Deep learning ; Gait analysis

Bioengineering and Biology:461 ; Biomechanics, Bionics and Biomimetics:461.3

Web of Science

1.Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen, Guangdong, Peoples R China
2.Univ Malaga, Dept Comp Architecture, Malaga, Spain
3.Univ Cordoba, Dept Comp & Numer Anal, Cordoba, Spain
4.Shenzhen Inst Artificial Intelligence & Robot Soc, Shenzhen, Guangdong, Peoples R China
5.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen, Guangdong, Peoples R China

Mu, Zihao,Castro, Francisco M.,Marin-Jimenez, Manuel J.,et al. iLGaCo: Incremental Learning of Gait Covariate Factors[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE,2020:1-8.