Break-Audio-Steganalysis-with-GAN

Towards Breaking Audio Steganalysis with Generative Adversarial Network

Prerequisites

1. Pytorch v0.4.1
2. CUDA preferred
3. Python v3.7

Model load

The chen.pkl is the trained weights of Chen's network and the best.pkl is the trained weights of our generator. They can be used directly.

Sample

The sample can be found in Google Drive.