An End-to-End 3D-CNN Framework for Silhouette-Based Human Gait Recognition
International Journal of Innovations in Science and Technology · 10.07.2026 · 1539
DOI: 10.33411/ijist/1953 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Innovations in Science and Technology, online erschienen am 10.07.2026. Verfasst von Asif Mehmood und 4 weiteren Autor:innen. Die Arbeit nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Human gait recognition (HGR) is a method of biometric identification that is widely adopted in fields of identification and authentication. Despite the fact that such a method is effective, it suffers from certain limitations caused by common variations, including illumination, carrying-condition, and walking-speed variations. In this paper, a new method called the 3D-CNN model is designed from scratch to overcome this problem. The proposed model uses silhouettes to obtain both spatial and temporal features. The proposed approach is based on some steps, such as acquisition of data, preprocessing, augmentation, designing of the 3D-CNN model, training of the model, and evaluation. The 3D-CNN model comprises four convolution blocks followed by a classification head. Experimental evaluation on the CASIA-C dataset demonstrated a validation accuracy of 98.48%, with a macro precision of 98.22%, macro recall of 97.60%, and macro F1-score of 97.53% by setting the number of epochs to 150. While accuracies of 95.87% and 96.52% were achieved using 100 and 200 epochs,respectively
Zitieren
APA
Mehmood, A., Rana, M. R. R., Shah, Z. M., Yousaf, M., & Yousuf, N. (2026). An End-to-End 3D-CNN Framework for Silhouette-Based Human Gait Recognition. International Journal of Innovations in Science and Technology, 1539. https://doi.org/10.33411/ijist/1953
Vancouver
Mehmood A, Rana MRR, Shah ZM, Yousaf M, Yousuf N. An End-to-End 3D-CNN Framework for Silhouette-Based Human Gait Recognition. International Journal of Innovations in Science and Technology. 2026:1539. doi:10.33411/ijist/1953
BibTeX
@article{Mehmood2026,
author = {Mehmood, Asif and Rana, Muhammad Rizwan Rashid and Shah, Zohaib Mehmood and Yousaf, Mehran and Yousuf, Naveed},
title = {An End-to-End 3D-CNN Framework for Silhouette-Based Human Gait Recognition},
journal = {International Journal of Innovations in Science and Technology},
year = {2026},
pages = {1539},
doi = {10.33411/ijist/1953}
}
EUVIMED