Performance comparison of colour correction and colour grading algorithm for medical imaging applications

Different types of image acquisition devices render different measure of colour depending on the specification of devices; even a same device will give different values of colours rendered, taking at certain duration of times. Most of the researches nowadays have attempted to solve these limitatio...

Full description

Bibliographic Details
Main Authors: Nur Diyana Kamarudin, Muhammad Azrae Yusof, Mohd Shahrizal Rusli, Ooi, Chia Yee, Syarifah Bahiyah Rahayu Syed Mansoor, Afiqah Mohd Azahari, Zuraini Zainol, Kamaruddin Abd Ghani, Siti Noormiza Makhtar
Format: Article
Language:English
Published: Penerbit Universiti Kebangsaan Malaysia 2021
Online Access:http://journalarticle.ukm.my/19079/
http://journalarticle.ukm.my/19079/1/21.pdf
Description
Summary:Different types of image acquisition devices render different measure of colour depending on the specification of devices; even a same device will give different values of colours rendered, taking at certain duration of times. Most of the researches nowadays have attempted to solve these limitations and the researches of colour correction algorithm has been evolved recently. Colour correction algorithm has been widely used in various fields such as food industry, medical imaging, forensic cyber applications, film industries etc. In medical imaging, researchers have considered colour correction as an essential part in their pre-processing step prior to diagnosis. There are various statistical methods in colour correction and colour grading algorithm being implemented nowadays and finding the best algorithm with high accuracy is non-trivial. This paper presents comparative analyses of colour correction techniques that combine colour correction and colour grading algorithm using conventional gamma correction, polynomial regression, and proposed polynomial regression with modified gamma Look-up Table (pgLUT). It has been observed that our proposed pgLUT colour correction algorithm outperformed the conventional methods by 16.5%.