VeriGenAI 1.0: Ανίχνευση Συνθετικών Εικόνων και Επαλήθευση Ηλικίας


Δημοσιευμένα: Φεβ 24, 2026
Λέξεις-κλειδιά:
GeAI Image Authenticity Verification Trustworthiness
Ryan Grissett
Epameinondas P Pliogkas
Stephanie Garay
Eleni Siamtanidou
Sevgi Grisset
Anna Podara
Loucas Protopappas
Iordanis Thoidis
Rigas Kotsakis
Lazaros Vrysis
Dimitrios Damopoulos
Περίληψη

The proliferation of generative AI poses a growing threat to digital identity systems, particularly in Know Your Customer (KYC) and underage verification workflows. This paper introduces VeriGenAI, a modular, privacy-first framework designed to detect AI-generated images in sensitive verification scenarios. Our system integrates facial landmark alignment, multimodal feature extraction, and transformer-based inference with counterfactual reconstruction using ComfyUI pipelines. Through an ensemble scoring model—combining GAN artifact detection, posture plausibility, age estimation, and metadata analysis—we achieve real-time judgments on image authenticity and demographic plausibility. The architecture supports air-gapped deployments and enables human-in-the-loop oversight. We further address challenges in latency, demographic bias, and adversarial robustness, offering transparent explainability features and forensic auditability. Beyond KYC and youth protection, VeriGen scales to applications in misinformation prevention, content moderation, and digital trust assurance. This work advances the frontier of trustworthy AI in identity validation through a secure, adaptable, and ethically grounded platform, while discusses the challenges of misusing such systems.

Λεπτομέρειες άρθρου
  • Ενότητα
  • 4ο Ελληνόφωνο Επιστημονικό Συνέδριο Εργαστηρίων Επικοινωνίας
Αναφορές
AU10TIX. (2025). Serial fraud monitor.
Bazarevsky, V., Kartynnik, Y., Vakunov, A., Tkachenka, A., & Grundmann, M. (2020, August). BlazePose: On-device real-time body pose tracking. Google AI Blog. https://ai.googleblog.com/2020/08/on-device-real-time-body-pose.html
Begum, M., Shorif, S. B., Uddin, M. S., Ferdush, J., Jan, T., Barros, A., & Whaiduzzaman, M. (2024). Image watermarking using discrete wavelet transform and singular value decomposition for enhanced imperceptibility and robustness. Algorithms, 17(1), 32. https://doi.org/10.3390/a17010032
Belhadi, A., Kamble, S. S., & Jabbour, C. J. C. (2023). Artificial intelligence and fraud detection. ResearchGate. https://www.researchgate.net/publication/357509652_Artificial_Intelligence_and_Fraud_Detection
Bradski, G. (2000). The OpenCV library. Dr. Dobb’s Journal of Software Tools. https://opencv.org/
Brunner, H. (2022). Piexif: Simplify EXIF handling with Python. GitHub. https://github.com/hMatoba/Piexif
Cao, Z., Simon, T., Wei, S.-E., & Sheikh, Y. (2017). Realtime multi-person 2D pose estimation using part affinity fields. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://openaccess.thecvf.com/content_cvpr_2017/papers/Cao_Realtime_Multi-Person_2D_CVPR_2017_paper.pdf
Dhariwal, P., & Nichol, A. (2021). Diffusion models beat GANs on image synthesis. arXiv Preprint arXiv:2105.05233. https://arxiv.org/abs/2105.05233
European Parliament and Council. (2024). Artificial Intelligence Act (EU) 2024/1689. Official Journal of the European Union. http://data.europa.eu/eli/reg/2024/1689/oj
Gulnazaki, G. (2025). Benchmarking counterfactual image generation. https://gulnazaki.github.io/counterfactual-benchmark/
Hu, E. J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, L., & Chen, W. (2021). LoRA: Low-rank adaptation of large language models. arXiv Preprint arXiv:2106.09685. https://arxiv.org/abs/2106.09685
Jeenath Laila, T., & Tamilpavai, G. (2023). FinBTech: Blockchain-based video and voice authentication system for enhanced security in financial transactions utilizing FaceNet512 and Gaussian mixture models. arXiv Preprint arXiv:2310.18668. https://arxiv.org/abs/2310.18668
LexisNexis Risk Solutions. (2024). Synthetic identity fraud. https://risk.lexisnexis.com/insights-resources/article/synthetic-identity-fraud
Lugaresi, C., Tang, J., Poulain, T., Chang, C., Yong, M., Lee, J., … & Kalantri, H. (2019). MediaPipe: A framework for building perception pipelines. arXiv Preprint arXiv:1906.08172. https://arxiv.org/abs/1906.08172
Mahmood, N., Ghorbani, N., Troje, N. F., Pons-Moll, G., & Black, M. J. (2019). AMASS: Archive of motion capture as surface shapes. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). https://amass.is.tue.mpg.de/
MDPI. (2025). Advancing GAN deepfake detection: Mixed datasets and residual artifacts. Applied Sciences, 15(2), 923. https://www.mdpi.com/2076-3417/15/2/923
Nguyen, T. H., Lee, J. H., & Tran, M. T. (2024). Social media authentication and combating deepfakes using semi-fragile watermarking. arXiv Preprint arXiv:2410.01906. https://arxiv.org/abs/2410.01906
Ollama. (2023). Ollama: Run large language models locally. GitHub. https://github.com/ollama/ollama
Reality Defender. (2025). How deepfakes exploit KYC verification systems. https://www.realitydefender.com/insights/how-deepfakes-exploit-kyc-verification-systems
Reuters. (2024, May 21). Europe sets benchmark for rest of the world with landmark AI laws. Reuters. https://www.reuters.com/world/europe/eu-countries-back-landmark-artificial-intelligence-rules-2024-05-21/
Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2022). High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). https://arxiv.org/abs/2112.10752
Rothe, R., Timofte, R., & Van Gool, L. (2015). DEX: Deep expectation of apparent age from a single image. In Proceedings of the IEEE International Conference on Computer Vision Workshops (ICCVW). https://doi.org/10.1109/ICCVW.2015.41
Saharia, C., Chan, W., Saxena, S., Li, L., Whang, J., Denton, E., … & Salakhutdinov, R. (2022). Photorealistic text-to-image diffusion models with deep language understanding. arXiv Preprint arXiv:2205.11487. https://arxiv.org/abs/2205.11487
Schlatt, V., Sedlmeir, J., Feulner, S., & Urbach, N. (2021). Designing a framework for digital KYC processes built on blockchain-based self-sovereign identity. arXiv Preprint arXiv:2112.01237. https://arxiv.org/abs/2112.01237
Smith, J., Doe, A., & Lee, K. (2025). Zero-shot synthetic image detector leveraging CLIP-based feature alignment. arXiv Preprint arXiv:2504.03765. https://arxiv.org/abs/2504.03765
Trend Micro. (2024). AI vs AI: DeepFakes and eKYC. Trend Micro. https://www.trendmicro.com/vinfo/us/security/news/cyber-attacks/ai-vs-ai-deepfakes-and-ekyc
Vaidya, A., & Awasthi, A. (2025). Zero-to-one IDV: A conceptual model for AI-powered identity verification. arXiv Preprint arXiv:2503.08734. https://arxiv.org/abs/2503.08734
Zhang, L., & Agrawala, M. (2023). Adding conditional control to text-to-image diffusion models. ControlNet. https://github.com/lllyasviel/ControlNet
Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://arxiv.org/abs/1801.03924
Zhang, Y., Zheng, Y., Wu, B., & Wang, Y. (2019). Detecting GAN-generated fake images using co-occurrence matrices. arXiv Preprint arXiv:1904.04068. https://arxiv.org/abs/1904.04068
Zhang, Z., Song, Y., & Qi, H. (2017). Age progression/regression by conditional adversarial autoencoder. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://susanqq.github.io/UTKFace/
Zhao, Y., & Wang, X. (2024). Optical image authentication and encryption scheme with computational ghost imaging and QR code. Applied Mathematical Modelling, 131, 1–14. https://doi.org/10.1016/j.apm.2024.03.017