PortraitProject

This repository contains the implementation of Light Maps for relighting portraits. The code is based on the paper LightMaps.pdf.

Steps for Execution:

1. Follow the Step 1 mentioned in DSINE GitHub to install the required dependencies and set up the necessary files.
2. Replace test_minimal.py with normals.py in the folder projects/dsine/.
3. Save the file light_maps.py in the same folder.
4. Save the input images in the folder projects/samples/img/.
5. Update the input image path in light_maps.py to point to your desired image. For example:

   file_path = os.path.abspath('samples/img/input_img.hdr')

6. Navigate to the folder projects/dsine/, and run the following command:

   python light_maps.py ./experiments/exp001_cvpr2024/dsine.txt

7. The output images are save in projects/samples/output/