RRDNet

Implementation of "Zero-Shot Restoration of Underexposed Images via Robust Retinex Decomposition, International Conference on Multimedia and Expo, 2020"

1. Prerequisites

• Python 3
• PyTorch >= 0.4.1
• PIL >= 6.1.0
• Opencv-python>=3.4

2. Run the project

There are two test underexposed images in test folder for demo.

Run

python3 pipline.py

and you will get result images in test folder.

You can change the path of input image in conf.py:

test_image_path = '/the/path/to/input/image'