Deepfake Detection with Metaheuristic Algorithms and Deep Features


Creative Commons License

KOÇAK A., ALKAN M.

INFORMATION TECHNOLOGY AND CONTROL, cilt.54, sa.4, 2025 (SCI-Expanded, Scopus)

Özet

The rapid development and spread of deepfake technology have posed serious threats to cybersecuri-ty, information security, privacy, and public safety; consequently, reliable detection of deepfake content has become a critical necessity. We present a metaheuristic-based hybrid approach that combines deep learning architectures with metaheuristic algorithms. In the study conducted on FaceForensics (FF++) and Celeb-DF datasets, four feature vectors extracted from Xception and ResNet50 architectures under-went a metaheuristic-based feature selection process incorporating Cuckoo Search Algorithm (CS), Suc-cess-History Based Adaptive Differential Evolution with Linear Population Size Reduction (L-SHADE), Particle Swarm Optimization (PSO) and Weighted Mean of Vectors Optimization Algorithm (INFO) algo-rithms. Sub-feature vectors were obtained through feature selection for each algorithm. The four feature vectors obtained from the architectures and sixteen sub-feature vectors generated after feature selection were classified using machine learning and deep learning methods. When comparing performance met-rics before and after feature selection, the INFO algorithm provided the highest performance across both datasets, achieving AUC values of 99.05% for the FF++ dataset and 99.01% for the Celeb-DF dataset. We believe that our comprehensive experimental study demonstrates more significant and effective results than existing methods