Text steganography Pytorch-Version

Summary

This repository contains a Pytorch reimplementation for previously prevailing text steganography methods based on LSTM language model.

The codebase is heavily insprired by the ADG-based steganography methods

Steganography methods

Huffman-coding based method (HC.py)

Arithmetic-coding based method (AC.py)

Patient-huffman based method (P-HC.py)

ADG-based method (ADG.py)

The implementation of the GAN-APD text steganography (2021 TDSC paper) is here!

Other Files

The code of training LSTM language model (train.py)

The code of architecture of LSTM Language model (lm.py)

The code of generating no-embedding text by well-trained LM with random sampling (Generated_Cover.py)

The code of Huffman algorithm (Huffman.py)

The other codes of these text-stegangography methods (utils.py)