Alternate Least Square and Root Polynomial Based Colour-Correction Method for High Dimensional Environment

The colours of a digital image rely not only on lighting conditions and the features of the capturing device but also on the surface qualities of the things included in the picture. The calculation of scene colorimetric from raw data remains an unresolved problem, particularly for digital photographs taken by digital image-capturing equipment under ambiguous lighting conditions. As a result, this work proposes an efficient and cost-efficient method for colour correction that combines Root Polynomial (RP) as well as Alternate Least Square (ALS) methodologies. Reducing errors within the reference picture and the target image is the suggested model’s main goal to raise the ultimate performance of the model. We then applied a combined ALS RP-based colour-correcting algorithm to the objective images to address this problem. To make colour coordinates easier to grasp, we additionally translated the example reference image as well as the target image into multiple different colour spaces such as LAB colour space, LUV colour space, and finally RGB. The proposed scheme is evaluated by the use of the Amsterdam Library of Object Images (ALOI) dataset and simulations are conducted using MATLAB software. Different performance matrices, such as Mean, Median, 95% Quantile, and Maximum Errors, are used to determine the simulated results. The outcome of these parameters in terms of various models shows that applying the suggested colour correction models results in the least amount of error difference between two images, indicating that colour transfer is accomplished smoothly.