White Balancing is the process of estimating the chromaticity(color) of the illuminant which was used to light the scene. The estimated color of the illuminant can then be used to generate a new image such that it has been taken under standard (D65) illuminant. This can be achieved through an independent gain regulation of the three color signals. This process results in reproducing a white object white (no color cast) irrespective of the color of the illuminant used. This ensures color constancy.
The goal of this task is to explore different variants of the classical AWB (Automatic White Balancing) algorithms currently being used and among which can be implemented in AXIOM Beta. This task is also for implementing the state of the art AWB algorithms for OpenCine which have better results but run time is slow and are computationally bit expensive.
Additional Information :
Color Balance : https://en.wikipedia.org/wiki/Color_balance
White Point : https://en.wikipedia.org/wiki/White_point
Tutorial on White Balance : https://www.cambridgeincolour.com/tutorials/white-balance.htm
AWB in Digital Photography : https://courses.cs.washington.edu/courses/cse467/08au/labs/l5/whiteBalance.pdf
AWB algorithm using regression trees : https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Cheng_Effective_Learning-Based_Illuminant_2015_CVPR_paper.pdf
AWB using CNNs : http://openaccess.thecvf.com/content_cvpr_2017/papers/Hu_FC4_Fully_Convolutional_CVPR_2017_paper.pdf
Survey on AWB algorithms : https://ieeexplore.ieee.org/document/5719167
To get in touch with any mentor check the [[ https://www.apertus.org/GSoC-2019-Mentor-Contact-List | Mentor Contact List]].