This repository contains the implementations of the evaluation metrics used in the paper "Perfectly Secure Steganography in Chat-Apps" by Marius Karstedt.
Gemini Code Assist was used in the development process of this implementation.
@misc{karstedt-2026-perfectly-secure-stego,
title = "Perfectly Secure Steganography in Chat-Apps",
author = "Karstedt, Marius",
month = jan,
year = "2026",
address = "Darmstadt, Germany",
url = "https://marius-karstedt.de/papers/perfectly-secure-steganography-in-chat-apps.pdf",
abstract = "This paper analyses various existing text steganography tools using a novel evaluation framework for the use in instant messaging (IM) applications. Unlike existing frameworks, the authors introduce metrics that evaluate tools in a real-world constrained environment. The paper aims to assess the suitability of various text steganography tools and their methodologies for the use in IM applications, with the goal of combatting censorship and encryption bans. The authors demonstrate, in this context, relying solely on statistical similarity (e.g., KL = 0) and a thematic coherence is insufficient to define a steganographic system as perfectly secure."
}