The rapid development of artificial intelligence has led to the emergence of deepfakes, which pose serious threats to information security and public trust in digital media. This study develops a facial deepfake detection system that integrates YOLOv11 for face detection and the Xception architecture for classifying real and manipulated faces. YOLOv11 successfully localized all facial regions in the tested dataset with high confidence scores. The Xception model achieved a testing accuracy of 90.10%, with a Recall of 97.11% for the Fake class and an AUC of 0.98. Visual explanation using Grad-CAM showed that the model focused on critical areas such as the forehead, temples, and face boundaries to detect manipulation artifacts. The system was implemented as a desktop application named "Snap Detector" and passed black-box testing. However, the average processing speed of 6.26 FPS on an NVIDIA T4 GPU indicates that further optimization is needed for real-time performance.
Fachril Akbar, Kurniawati· JOURNAL OF APPLIED INFORMATI...· 0 citations
Horizontal Privilege Escalation (HPE) is a challenging access control vulnerability because it is performed by authenticated users whose activities closely resemble legitimate behavior, making detection difficult for conventional security mechanisms. Most existing studies rely on single-log analysis, which limits their ability to capture behavioral patterns distributed across multiple log sources. This study presents a comparative analysis of three supervised machine learning algorithms—Support Vector Machine (SVM), Random Forest, and XGBoost—for HPE detection in web applications using a multi-log correlation framework. The proposed approach integrates authentication, access, and activity logs to construct a behavior-oriented feature representation. Experiments were conducted on the publicly available Access-Log-Anomaly-Detection-Dataset containing 5,000 labeled records. The models were trained using stratified five-fold cross-validation and evaluated using Accuracy, Precision, Recall, F1-score, and ROC-AUC. Experimental results show that all models achieved comparable performance, with accuracies ranging from 84.0% to 85.1% and ROC-AUC values above 0.92. Random Forest achieved the highest Accuracy (85.1%) and F1-score (0.8216), while SVM obtained the highest Recall (0.8378) and XGBoost the highest Precision (0.8203). These findings demonstrate that multi-log correlation combined with behavior-oriented feature engineering provides an effective foundation for machine learning-based HPE detection in web applications.
Riska Dami Ristanto, Ahmad Fauzan, Kurniawati et al.· MATEC Web of Conferences· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
