A python project that takes an image of an ishihara test (colorblindness test) and outputs the recognised number!
The project works by applying advanced image procesing techniques to extract the digits from the image (works for 1 and 2 digit numbers). After the digits are extracted, they are recognised by a custom made convolutional neural network trained on 1400 preprocesed images of ishihara test
Big thanks to the creator of the awesome dataset used for training the model: dataset
An example image of an ishihara test used to diagnose color blindness
The digit extracted from the ishihara test after image preprocesing