Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal Academia Open
Zaid Derea
College of Computer Science and Information Technology, Wasit University, Wasit 52001
Medicine & Pharmacology Public Health

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
GAN and DINOv2 Framework for Robust Cross-Condition Gait Recognition Zaid Derea
Academia Open Vol. 11 No. 1 (2026): June
Publisher : Universitas Muhammadiyah Sidoarjo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21070/acopen.11.2026.13411

Abstract

General Background: Gait recognition is a remote, non-invasive biometric widely used in forensics, surveillance, and security systems. Specific Background: Deep learning has advanced gait analysis, yet CNN-based approaches struggle with temporal coherence, cross-view variation, and degraded silhouettes. Knowledge Gap: Existing studies typically separate gait synthesis and recognition, with limited use of self-supervised transformers and insufficient joint evaluation of generative quality and identification performance. Aims: This study proposes an integrated framework combining multi-GAN gait reconstruction, DINOv2 vision transformer feature extraction, and CNN-based identity classification. Results: StyleGAN2 produced the most realistic silhouettes (PSNR 31.2 dB, SSIM 0.925, FID 18.3), while DINOv2 yielded highly separable 768-dimensional features, leading to 98.3% classification accuracy across varied walking conditions and strong clustering metrics (NMI 0.891, ARI 0.847). Novelty: The work unifies generative gait synthesis, transformer-guided spatiotemporal representation, and comprehensive evaluation within a single pipeline. Implications: The framework supports reliable biometric verification in forensic investigation, surveillance monitoring, rehabilitation assessment, and real-time security deployment under clothing, view, and carrying variations. Highlights: Joint GAN synthesis and transformer features within one gait recognition pipeline. High-quality silhouette reconstruction linked to superior identity separability. Stable recognition (>94%) under clothing, view, speed, and carrying variations. Keywords: Gait Recognition, GAN, DINOv2, Biometric Identification, Computer Vision
Improving human gait recognition using a directed GAN in digital forensics : Peningkatan pengenalan gerakan manusia menggunakan GAN terarah dalam forensik digital Zaid Derea
Academia Open Vol. 11 No. 1 (2026): June
Publisher : Universitas Muhammadiyah Sidoarjo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21070/acopen.11.2026.13545

Abstract

General Background: Gait recognition has emerged as a non-invasive biometric technique widely applied in forensic analysis, surveillance, and security systems. Specific Background: Recent advances in deep learning, particularly Generative Adversarial Networks and vision transformers, have enabled improved gait synthesis and feature representation, yet challenges remain in maintaining identity consistency and evaluation reliability. Knowledge Gap: Existing studies often address gait synthesis or recognition independently and lack unified frameworks combining generative quality assessment with robust feature extraction. Aims: This study proposes a comprehensive gait recognition framework integrating GAN-based gait enhancement, DINOv2 vision transformer feature extraction, and CNN-based identity verification. Results: Experimental evaluation on CASIA-B, OU-ISIR, and TUM-GAID datasets shows that StyleGAN2 achieves superior gait reconstruction quality, while DINOv2 provides highly discriminative spatio-temporal features, resulting in a classification accuracy of 98.3% and strong clustering separability. Novelty: The framework uniquely combines competitive GAN architectures with self-supervised transformer features and multi-level evaluation metrics within a single system. Implications: The proposed approach supports reliable gait recognition for biometric verification, forensic investigation, healthcare monitoring, and surveillance applications. Keywords: Gait Recognition, Generative Adversarial Networks, DINOv2, Vision Transformers, Biometric Verification Key Findings Highlights: High-fidelity gait synthesis was achieved using advanced generative architectures. Self-supervised transformer features provided strong identity separability. The integrated framework demonstrated robust performance across diverse conditions.
Co-Authors