LeonOrou/ML_Image_Complementer

ML Project from 2nd semester AI studies for predicting missing pixels in image files Contains a Convolutional Neural Network (CNN) that is trained to predict missing pixels from images. The missing pixels are stand-alone seperated by an x- and y-spacing which have for the training values between 2 and inlcuding 6 pixels. The missing pixels are starting at an x- and y-offset which have for training values between 0 and 8 pixels. The final accuracy measured as the root mean squared error of the predicted RGB values were around 18.42 pixels.