Related Experiment Video
Updated: Mar 26, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.2K
Multi-Scale Fusion for Improved Localization of Malicious Tampering in Digital Images
Summary
This study introduces a multi-scale analysis to improve image tampering localization. By fusing results from different window sizes, it achieves higher resolution and reliability in detecting manipulated image regions.
Area of Science:
- Digital Image Forensics
- Computer Vision
- Information Security
Background:
- Sliding window analysis is common for image tampering localization.
- Current methods suffer from low resolution or unreliable classification due to window size limitations.
Purpose of the Study:
- To investigate a multi-scale analysis approach for enhancing image tampering localization.
- To fuse multiple candidate tampering maps for improved resolution and reliability.
Main Methods:
- Proposed three novel multi-scale fusion techniques.
- Utilized mode-based first digit features for distinguishing compression types.
- Evaluated performance against various reference strategies.
Main Results:
- Multi-scale fusion successfully combined benefits of small and large window analyses.
- Achieved improved tampering localization performance.
- Demonstrated enhanced reliability and resolution in detection maps.
Conclusions:
- The proposed fusion strategies offer a significant improvement over traditional window-based methods.
- Multi-scale analysis is a viable approach for robust passive image authentication.
- This technique enhances the accuracy of identifying manipulated image regions.