Impersonation in the Digital Age: A Comparative Review of Detection Techniques Against Deepfakes

@article{1710585,
      author = {Ochi Victor Chukwudi, Tochukwu Chinecherem Nnabuike},
      title = {Impersonation in the Digital Age: A Comparative Review of Detection Techniques Against Deepfakes},
      journal = {Iconic Research And Engineering Journals},
      year = {2025},
      volume = {9},
      number = {3},
      pages = {648-651},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1710585.pdf},
      abstract = {Deepfakes?AI?generated or manipulated audio?visual content?pose a growing risk to identity verification, trust, and security. This comparative review synthesizes state?of?the?art detection techniques across visual, audio, and multimodal pipelines, focusing on performance, generalization, robustness, and operational considerations (latency, cost, and privacy). We analyze benchmark datasets (e.g., FaceForensics++, DFDC, FakeAVCeleb, WildDeepfake, ASVspoof 2021) and evaluation metrics (AUC, EER, F1) and compare classical, deep, and hybrid approaches. We also examine adversarial pressures?including compression, unseen manipulations, and cross?domain shifts?and outline practical integration patterns for KYC/AML, exam proctoring, and access control. The review concludes with an engineering blueprint for a multimodal detection stack, emphasizing human?in?the?loop triage and risk scoring.},
      keywords = {Deepfakes, Impersonation Detection, Multimodal Biometrics, Audio Spoofing, Media Forensics, Identity Verification, KYC, Liveness Detection, Robustness, Generalization},
      month = {September},
  }