Real-world actor-based image steganalysis via classifier inconsistency detection

2023-08-29

ARES

AI Summary

This paper addresses the cover source mismatch (CSM) challenge in practical steganalysis—where training and testing covers originate from different distributions. We propose the first actor-level classifier inconsistency detection (DCI) framework for steganographer attribution, reframing CSM not as mere nuisance but as a quantifiable, decision-relevant diagnostic signal. Our method enables three-way classification: “innocent,” “guilty,” or “discard due to excessive CSM.” It integrates EfficientNet-based feature extraction, a novel DCI prediction mechanism that measures inter-classifier disagreement under CSM, and a gradient boosting machine (GBM) classifier. Evaluated under high-CSM conditions, our approach achieves stable accuracy exceeding 80%, outperforming existing baselines by a significant margin. This work establishes a new paradigm for steganographic behavior attribution under source distribution mismatch.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Adversarial Attacks & Robustness
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Security and Privacy: Large-scale security measurements
• Responsible Web: Measurement, analysis, and circumvention of Web censorship

Abstract

In this paper, we propose a robust method for detecting guilty actors in image steganography while effectively addressing the Cover Source Mismatch (CSM) problem, which arises when classifying images from one source using a classifier trained on images from another source. Designed for an actor-based scenario, our method combines the use of Detection of Classifier Inconsistencies (DCI) prediction with EfficientNet neural networks for feature extraction, and a Gradient Boosting Machine for the final classification. The proposed approach successfully determines whether an actor is innocent or guilty, or if they should be discarded due to excessive CSM. We show that the method remains reliable even in scenarios with high CSM, consistently achieving accuracy above 80% and outperforming the baseline method. This novel approach contributes to the field of steganalysis by offering a practical and efficient solution for handling CSM and detecting guilty actors in real-world applications.

Problem

• Anomaly Detection
• Image Analysis
• Behavior Recognition

Innovation

• EfficientNet Neural Network
• Gradient Boosting Machine
• Image Source Variation Handling