Perceptual Hashing Using Pretrained Vision Transformers

被引:0
|
作者
De Geest, Jelle [1 ]
De Smet, Patrick [2 ]
Bonetto, Lucio [2 ]
Lambert, Peter [1 ]
Van Wallendael, Glenn [1 ]
Mareen, Hannes [1 ]
机构
[1] Univ Ghent, Imec, Dept Elect & Informat Syst, Technol Pk Zwijnaarde 122, B-9052 Ghent, Belgium
[2] Natl Inst Criminalist & Criminol NICC, Vilvoordsesteenweg 100, B-1120 Brussels, Belgium
来源
2024 IEEE GAMING, ENTERTAINMENT, AND MEDIA CONFERENCE, GEM 2024 | 2024年
关键词
Perceptual Hashing; Vision Transformer; Image Forensics;
D O I
10.1109/GEM61861.2024.10585453
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
The rapid evolution of digital image circulation has necessitated robust techniques for image identification and comparison, particularly for sensitive applications such as detecting Child Sexual Abuse Material (CSAM) and preventing the spread of harmful content online. Traditional perceptual hashing methods, while useful, fall short when exposed to some common image transformations, or when images are doctored to avoid detection, rendering them ineffective for nuanced comparisons. Addressing this challenge, this paper introduces a novel pretrained vision transformer artificial intelligence (AI) model approach that enhances the robustness and accuracy of perceptual hashing. Leveraging a pretrained Vision Transformer (ViT-L/14), our approach integrates visual and textual data processing to generate feature arrays that represent perceptual image hashes. Through a comprehensive evaluation using a dataset of 50,000 images, we demonstrate that our method offers significant improvements in detecting similarities for certain complex image transformations, aligning more closely with human visual perception than conventional methods. While our method presents certain initial drawbacks such as larger hash sizes and high computational complexity, its ability to better handle perceptual nuances presents a forward step in the realm of image forensics. The potential applications of this research extend to law enforcement, digital media management, and the broader domain of content verification, setting the stage for more secure and efficient digital content analysis.
引用
收藏
页码:19 / 24
页数:6
相关论文
共 50 条
  • [41] A Robust Video Identification Framework using Perceptual Image Hashing
    Vega, Francisco
    Medina, Jose
    Mendoza, Daniel
    Saquicela, Victor
    Espinoza, Mauricio
    2017 XLIII LATIN AMERICAN COMPUTER CONFERENCE (CLEI), 2017,
  • [42] Perceptual Audio Hashing using RT and DCT in wavelet domain
    Li, Jinfeng
    Wu, Tao
    2015 11TH INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE AND SECURITY (CIS), 2015, : 363 - 366
  • [43] Perceptual Audio Hashing Functions
    Hamza Özer
    Bülent Sankur
    Nasir Memon
    Emin Anarım
    EURASIP Journal on Advances in Signal Processing, 2005
  • [44] Supervised perceptual image hashing using collective matrix factorization
    Du, Ling
    Wang, Ziwei
    Su, Hua
    JOURNAL OF ELECTRONIC IMAGING, 2024, 33 (06)
  • [45] Perceptual hashing for color images
    Li, Xinran
    Qin, Chuan
    Yao, Heng
    Li, Jian
    JOURNAL OF ELECTRONIC IMAGING, 2021, 30 (06)
  • [46] Perceptual audio hashing functions
    Özer, H
    Sankur, B
    Memon, N
    Anarim, E
    EURASIP JOURNAL ON APPLIED SIGNAL PROCESSING, 2005, 2005 (12) : 1780 - 1793
  • [47] Reversible Multipurpose Watermarking Algorithm Using ResNet and Perceptual Hashing
    Jiang, Mingfang
    Yang, Hengfu
    JOURNAL OF INFORMATION PROCESSING SYSTEMS, 2023, 19 (06): : 756 - 766
  • [48] Perceptual Image Hashing Using Feature Fusion of Orthogonal Moments
    Li, Xinran
    Wang, Zichi
    Feng, Guorui
    Zhang, Xinpeng
    Qin, Chuan
    IEEE TRANSACTIONS ON MULTIMEDIA, 2024, 26 : 10041 - 10054
  • [49] Publicly Evaluatable Perceptual Hashing
    Gennaro, Rosario
    Hadaller, David
    Jafarikhah, Tahereh
    Liu, Zhuobang
    Skeith, William E.
    Timashova, Anastasiia
    APPLIED CRYPTOGRAPHY AND NETWORK SECURITY (ACNS 2020), PT II, 2020, 12147 : 436 - 455
  • [50] Perceptual audio hashing functions
    Özer, H. (hozer@uekae.tubitak.gov.tr), 1780, Hindawi Publishing Corporation (2005):