Daltonizer: A CNN-based Framework for Monochromatic and Dichromatic Color-Blindness

Color blindness is a condition in which an individual has difficulty distinguishing between certain colors or perceiving them correctly. Human eyes consist of color-delicate cells that are responsible for providing the information required by the brain to produce the ‘discerned’ color in an individual’s visible spectrum. These cells that are sensitive to different colors respond differently to various light wavelengths. The cells can’t be rectified in any way. This suggests color blindness cannot be cured. The primary focus of this research is on the following types of color deficiencies including complete color-blindness and dichromatic color-blindness—red-green, blue-yellow. In this paper, the authors use Image-Processing-Techniques (IPT) to make Ishihara-Test-Images (a test to identify partial and total color deficiency) perceptible to people with color-vision difficulties. The use of advanced image processing techniques (IPT) and mathematical operations on the pixels of an image can solve vision defects. To demonstrate this, a convolutional neural network (CNN) was designed and trained on the MNIST dataset to enhance color perception for individuals with achromatopsia.