Important

The dataset "SONY_IMX135" produced by INTEL-TAU is not included in the repository due to its size of 2.33 GB. It is avaliable to download here.

Spatially-Adaptive Log-Chroma White Balance

A Spatially-Adaptive Log-Chroma White Balance (SALC-WB) algorithm on raw images. Traditional methods, such as Gray World and Shades-of-Gray, compute global statistical estimations that often fail in locally varied lighting. This pipeline combines local statistical estimation in log-chroma space with a Shades-of-Gray prior. Unlike existing methods that require training or complex optimizations, is a simple and efficient AWB algorithm that recovers accurate luminance and reduces local artifacts.

How to run

1. Clone the repository

git clone https://github.com/akrossu/salc-wb.git
cd salc-wb

2. Download and Extract the contents of the Sony_IMX135 zip into the root folder of the project

.
├── Sony_IMX135/
|   ├── field_3_cameras/
|   ├── lab_printouts/
|   └── lab_realscene
├── lib/
├── Paper/
├── Presentation/
├── README.md
├── requirements.txt
├── salc-wb.py
└── tests.py

4. Create a virtual python environment

5. Install the required libraries using pip install -r requirements.txt

6. Run salc-wb.py

Note: debug is default enabled, but can be set to false