💧 WMSuite: Image Watermark Processing Toolkit

WMSuite is a lightweight Python toolkit that integrates several existing neural-network–based invisible watermarking algorithms.
The goal of this project is not to propose new watermarking methods, but to provide a simple and unified interface for:

• embedding watermarks into images from a specified folder, and
• extracting watermarks from images in a specified folder.

Currently, WMSuite includes support for 6–7 publicly available watermarking models, with a consistent configuration and execution workflow to make experimentation easier.

🚀 Quick Start Guide

1. Environment Setup

To set up the project environment, follow the steps below.

conda create -n <env_name> python=3.10

Activate the newly created environment.

conda activate <env_name>

Install all required Python dependencies from the provided requirements file.

pip install -r requirements.txt

2. Model Weights Configuration 📥

Several neural-network-based watermarking algorithms in this project require pretrained model weights.
Please follow the instructions below to download the corresponding weights and place them in:

• DwTDCT, RivaGAN

DwTDCT is a classical, non–neural-network watermarking algorithm, while RivaGAN is a neural watermarking method.
Both algorithms are integrated in this project by directly calling the Python API provided by: https://github.com/ShieldMnt/invisible-watermark.git

• HiDDeN

Download from: https://github.com/ando-khachatryan/HiDDeN.git

• StegaStamp

Download from: https://github.com/ningyu1991/ArtificialGANFingerprints.git

• Stable Signature

Download from: https://github.com/facebookresearch/stable_signature

• Vine

This watermarking algorithm does not require manual weight preparation.
When invoked for the first time, it will automatically download the required model weights from HuggingFace.

📁 Example Directory Structure (after downloading all weights)

algorithms/
└── checkpoints/
	├── hidden/
	│   ├── combined-noise--epoch-400.pyt
	│   ├── crop-epoch-300.pyt
	│   └── no-noise--epoch-400.pyt
	├── stable_signature/
	│   ├── dec_48b_whit.torchscript.pt
	│   ├── sd2_decoder.pth
	│   └── v2-1_512-ema-pruned.ckpt
	└── stegastamp/
		├── AFHQ_cat2dog_256x256_decoder.pth
		└── AFHQ_cat2dog_256x256_encoder.pth

---

3. Usage

WMSuite provides two basic functionalities: watermark embedding and watermark extraction.
You can run them directly using the provided shell scripts:

To embed watermarks into all images within a specified folder, run:

bash emb.sh

To extract watermarks from images in a specified folder, run:

bash extract.sh

🙏 Acknowledgements

This project integrates implementations and pretrained models from several existing watermarking algorithms.
We would like to acknowledge and thank the authors of these repositories.