Skip to content

Rich Model Steganalysis (SRM)

Definition

A feature-extraction approach that computes hundreds of statistical features from pixel-residual co-occurrence matrices and feeds them to an ensemble classifier such as a Fisher Linear Discriminant.

Feature source
Pixel-residual co-occurrence matrices
Feature count
Hundreds of statistical features
Example classifier
Fisher Linear Discriminant ensemble

Common questions

How does an SRM-based classifier decide if an image contains hidden data?+

It computes co-occurrence statistics from pixel-residual filters across the image, compares the resulting feature vector against patterns learned from labelled cover and stego image sets, and an ensemble classifier such as Fisher Linear Discriminant votes on the classification.

Why is an ensemble classifier preferred for rich-model features?+

The feature space is very high dimensional relative to typical training set sizes, and training many weak classifiers on random feature subsets and combining their votes reduces overfitting compared to training one classifier on the full feature vector.

Related terms

Blind Steganalysis
Detection without prior knowledge of the embedding algorithm. A classifier trained on images with and without payloads generalises across multiple tools.
Embedding Rate
The payload size divided by the carrier capacity, usually expressed as bits per pixel. Detection difficulty decreases sharply at low embedding rates...
False Positive Rate (FPR)
The proportion of true non-matches (or true negatives) that a classification method declares as matches (or positives). Also called the type I...
Steganalysis
The forensic discipline of detecting the presence of hidden data in a carrier file. Steganalysis uses statistical tests (chi-squared, RS analysis, sample...
Targeted Steganalysis
Detection methods designed against a specific steganography tool or algorithm. Effective when the tool is known but fails against novel or unknown...

Explained in

Your journey to becoming a forensic professional starts here.

Practice with mock tests, learn from structured notes, and get your questions answered by a global forensic community, all in one place.