Related Experiment Video
Updated: Oct 22, 2025
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
43.1K
A Detection Method of Operated Fake-Images Using Robust Hashing
Miki Tanaka1, Sayaka Shiota1, Hitoshi Kiya1
1Department of Computer Science, Tokyo Metropolitan University, 6-6 Asahigaoka, Tokyo 191-0065, Japan.
Journal of Imaging
|August 30, 2021
Summary
This study introduces a novel robust hashing method for detecting fake or tampered images, even after compression and resizing. The technique offers improved accuracy compared to existing methods, especially when a reference image is available.
Area of Science:
- Computer Science
- Digital Forensics
- Image Processing
Background:
- Social media platforms often alter uploaded images through recompression and resizing.
- Conventional fake image detection methods lack robustness against these common image manipulations.
Purpose of the Study:
- To develop a novel, robust method for detecting fake and tampered images.
- To address the limitations of existing methods against image compression and resizing.
Main Methods:
- Utilized a robust hashing method for fake/tampered image detection.
- Employed hash values from both reference and query images for detection.
- Introduced the use of original hash codes for enhanced comparison.
Main Results:
- The proposed method demonstrates superior robustness in detecting fake images compared to conventional techniques.
- Achieved state-of-the-art performance on various datasets, including those with GAN-generated fake images.
- Successfully detected distortions caused by image compression and resizing.
Conclusions:
- The novel robust hashing approach effectively detects fake images despite common manipulations.
- This method offers a significant advancement for digital forensics and image authenticity verification.
- Practical applications include monitoring synthetic images in commercial contexts.
