Application of Benford's Law in Deepfake Image Detection

被引：0
作者：
Nikitenkova, Svetlana P. [1 ]
机构：
[1] Lobachevsky State Univ Nizhny Novgorod, Nizhnii Novgorod, Russia
来源：
VESTNIK TOMSKOGO GOSUDARSTVENNOGO UNIVERSITETA-UPRAVLENIE VYCHISLITELNAJA TEHNIKA I INFORMATIKA-TOMSK STATE UNIVERSITY JOURNAL OF CONTROL AND COMPUTER SCIENCE | 2023年 / 64期
关键词：
AI-synthesized image; Benford's law; entropy; Kullback-Leibler divergences; Random Forest method; CLASSIFICATION; DIGIT;
D O I：
10.17223/19988605/64/13
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
The development and improvement of deepfake detection technologies is one of the priority areas for ensuring social and biometric security. The main goal of the work is to test the application of the well-known Benford's law as a detection tool for images generated by GAN. The proposed approach is based on the analysis of power spectrum and Shannon entropy of image. The effectiveness of the proposed method was tested on datasets generated by StyleGAN2 and StyleGAN3 neural networks. The proposed method does not require large computing power.
引用
收藏
页码：128 / 137
页数：10
相关论文
共 33 条
  • [1] Acebo E., 2005, P 1 EUR C COMP AESTH, P169
  • [2] Classification of image distortion based on the generalized Benford's law
    Al-Bandawi, Hussein
    Deng, Guang
    [J]. MULTIMEDIA TOOLS AND APPLICATIONS, 2019, 78 (18) : 25611 - 25628
  • [3] [Anonymous], GitHub-NVlabs/FFHQ-dataset
  • [4] [Anonymous], Stylegan3-dataset
  • [5] [Anonymous], scikit-learn: Machine Learning in Python-scikit-learn 1.1.1 Documentation
  • [6] [Anonymous], GitHub-NVlabs/StyleGAN2-dataset
  • [7] Benford F., 1938, P AM PHILOS SOC, V78, P551
  • [8] On the use of Benford's law to detect GAN-generated images
    Bonettini, Nicolo
    Bestagini, Paolo
    Milani, Simone
    Tubaro, Stefano
    [J]. 2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2021, : 5495 - 5502
  • [9] Evading Deepfake-Image Detectors with White- and Black-Box Attacks
    Carlini, Nicholas
    Farid, Hany
    [J]. 2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW 2020), 2020, : 2804 - 2813
  • [10] Corvi R, 2022, Arxiv, DOI arXiv:2211.00680
1234