Skip to content

Latest commit

 

History

History
90 lines (55 loc) · 2.6 KB

README.md

File metadata and controls

90 lines (55 loc) · 2.6 KB

Copy-Move Forgery Detection and Question Answering for Remote Sensing Image

This is the initial version of the RS-CMQA dataset, RS-CMQA-B dataset and Copy-Move Forgery Awareness Framework (CMFAF).

2024.9.5. initial version

2024.12.15. Updated code and dataset links

Installation

python >=3.10
conda create -n tamper python=3.10
conda activate tamper
pytorch

install pytorch

# e.g. CUDA 11.8
# with conda
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
# with pip
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
Install Packages
pip install -r requirements.txt

Download Datasets

  • Datasets V1.0 is released at Baidu Drive (2024.9.5). Available for download.

    Dataset v1 only includes copy-move forgery

  • Datasets V2.0 is released at Baidu Drive (2024.10.11). Will be available for download after the paper is officially accepted.

    Dataset v2 includes copy-move and blurring tamper. For blurring tamper, the tampered region and the source region are treated as the same region

  • TBD: The high-quality, annotated manually dataset will be released before March, 2025

  • Dataset Directory: datasets/

  • Dataset Subdirectory: datasets/JsonFiles/, datasets/JsonFilesBalanced/, datasets/image/, datasets/source/, datasets/target/, datasets/background/

Download pre-trained weights

Download clip-b-32 weights from Hugging Face

  • Clip Directory: models/clipModels/openai_clip_b_32/

Download U-Net weights from Github

  • U-Net Directory: models/imageModels/milesial_UNet/

Start Training

python main.py
  • Modify the experiment settings and hyperparameters in src/config.py

Data Examples

数据集

Citation

@article{zhang2024copymove,
    title={Copy-Move Forgery Detection and Question Answering for Remote Sensing Image}, 
    author={Z. Zhang and E. Zhao and Z. Wan and J. Nie and X. Liang and L. Huang},
    journal={arXiv preprint arXiv:2412.02575},
    year={2024},
}

License

CC BY-NC-SA 4.0

All images and their associated annotations in Global-TQA can be used for academic purposes only, but any commercial use is prohibited.