Enhanced Multi Attention-Net with Frequency-Aware Squeezed Residual Blocks for Robust Face Anti-Spoofing
International Journal of Engineering & Extended Technologies Research · 28.03.2026 · 08 (02)
DOI: 10.15662/ijeetr.2026.0802075 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Engineering & Extended Technologies Research, online erschienen am 28.03.2026. Verfasst von Janani G K und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Face recognition systems are widely deployed in authentication, surveillance, and financial applications. However, these systems remain highly vulnerable to presentation attacks such as printed photos, replayed videos, and 3D masks. Existing CNN-based anti-spoofing approaches often suffer from poor cross-dataset generalization due to environmental variations and unseen spoof patterns. This paper proposes an Enhanced Multi-Attention Network with Frequency-Aware Squeezed Residual Blocks (EMAN-FASRB) to improve robustness against face spoofing attacks. The proposed architecture integrates spatial, channel, and frequency attention mechanisms with lightweight residual learning to capture both spatial textures and frequency-domain spoof artifacts. Experiments conducted on benchmark datasets including CASIA-FASD, Replay-Attack, and MSU-MFSD demonstrate improved performance in terms of EER, HTER, ACER, and AUC. Results confirm superior cross-dataset generalization and computational efficiency compared to baseline CNN models
Zitieren
APA
K, J. G., N, J. S., A, N., & M, P. B. (2026). Enhanced Multi Attention-Net with Frequency-Aware Squeezed Residual Blocks for Robust Face Anti-Spoofing. International Journal of Engineering & Extended Technologies Research, 08(02). https://doi.org/10.15662/ijeetr.2026.0802075
Vancouver
K JG, N JS, A N, M PB. Enhanced Multi Attention-Net with Frequency-Aware Squeezed Residual Blocks for Robust Face Anti-Spoofing. International Journal of Engineering & Extended Technologies Research. 2026;08(02). doi:10.15662/ijeetr.2026.0802075
BibTeX
@article{K2026,
author = {K, Janani G and N, Jaya Shree and A, Nagaraj and M, Parisha Beham},
title = {Enhanced Multi Attention-Net with Frequency-Aware Squeezed Residual Blocks for Robust Face Anti-Spoofing},
journal = {International Journal of Engineering & Extended Technologies Research},
year = {2026},
volume = {08},
number = {02},
doi = {10.15662/ijeetr.2026.0802075}
}
EUVIMED