Color transformation for protanopia color vision deficiency using integration of image processing and artificial neural network

Abstract

Color blindness deficiency is inability to distinguish colors with each other. Nowadays, the individual who are not being able to recognize color may be crucial in some day life situation because many common activities depend on signals with color-coded such as road sign, traffic light, electric wire, resistor and many more. There are many forms of color blindness such Monochromacy (total color blindness), Dichromacy (Red/ Green/Blue blindness) and Trichromacy and etc. Most types of defective color blindness can be classified into two categories which are green color defective and red color defective. The objective of this project is to improve the ability of color discrimination for Protanopia which a type of dichromacy where the patients does not naturally develop red color or Long wavelength cones in their eyes. This project proposed a method using image processing to improve the ability of color discrimination for Protanopia as well as adjusting images such that a person suffering from Protanopia is able perceive image detail and color dynamics. This method is first developed by simulating an image through the eyes of a person suffering from protanopia by converting RGB space to LMS (long, medium, short) color space based on cone response and then modifies the response of the deficient cones. The linear multiplication matrix is derived by referring to CIE color matching functions. ANN is then set up by using the input/output from matrix conversion. For this research, the ANN is introduced to reduce simulation time in image processing. The transformation technique used is RGB Color Contrasting where this step is to enhance contrast between red and green which in general, make green pixels appear to be bluer. Based on the result, the objectives are successfully achieved. ANN gives the minimum computational time than conventional matrix conversion which is 36% increment. The changes of the image drastically for both color blind and non-color blind viewers. The result shows that the reds become redder and greens become greener from the image before being adjusted.