Enhancer - Image Enhancement Demo

This is an offline demo of Enhancer Image Enhancement.

1. Requirements

The basic requirements for this demo are:

• Ubuntu System
• Python 3+

2. Run Demo

We recommend creating a new conda virtual environment to run this demo, as follow:

1. Clone this repository:

   git clone https://github.com/ZHKKKe/Harmonizer.git
   cd Harmonizer
   

2. Download the pre-trained model enhancer.pth from this link and put it into the folder Harmonizer/pretrained/.

3. Create a conda virtual environment named harmonizer (if it doesn't exist) and activate it. Here we use python=3.8 as an example:

   conda create -n harmonizer python=3.8
   source activate harmonizer
   

4. Install PyTorch and the required python dependencies (please make sure your CUDA version is supported by the PyTorch version installed). In the root path of this repository, run:

   pip install -r src/requirements.txt
   

5. Execute the demo code in the root path of this repository, as:

   python -m demo.image_enhancement.run \
          --example-path ./demo/image_enhancement/example
   

   where ./demo/image_enhancement/example contains a sample we provided.
   If you want to test your own samples, please refer to the folder ./demo/image_enhancement/example to prepare the original image.

6. Check the image enhancement results in the folder: ./demo/image_enhancement/example/enhanced.