PatchRLGAN

Generative adversarial network with reinforcement learning for adaptive patch selection in medical and microscopy image denoising.

---

Usage

Setup Environment

1. Clone GitHub repository:

   git clone https://github.com/joaolcguerreiro/PatchRLGAN.git
   cd PatchRLGAN
   

2. Create environment:

   conda create -n patchrlgan python=3.10.11
   conda activate patchrlgan
   

3. Install Python dependencies (Python 3.10 is recommended):

   pip install -r requirements.txt
   

---

Train

1. Navigate to configs/config.yaml and adjust configuration parameters needed.

2. Navigate to src/scripts/.

3. Run script:

   python train.py
   

---

Test

1. Navigate to experiments/ and copy experiment path.

2. Navigate to src/scripts/.

3. Run script:

   python test.py --exp=<experiment_path> --epoch=<epoch>
   

Note: Results are in <experiment_path>/results.

---

License

This project is licensed under MIT license. See LICENSE for details.