Counter Histogram-Based Forensics using Mean Structural Similarity Index Metric
Exploiting mathematical properties of the SSIM index to counter histogram-based image forensics techniques.
Authors: R. Kazemi, A. Amini, B. Khodabandeh, M. Alikhani
Venue: SIAM Journal on Imaging Sciences, 2025
Histogram-based forensic methods detect image manipulation by identifying statistical anomalies in pixel value distributions. Counter-forensics aims to understand the limits of these detectors by designing images that evade detection while preserving visual quality.
This work analyzes the mathematical properties of the Structural Similarity Index (SSIM) and uses them to construct counter-forensic images that are both visually indistinguishable from the original and statistically undetectable by histogram-based forensic tools.
This was a voluntary research contribution during my undergraduate studies.
