This model implementation of CNN is used to predict the missing pixels of input an image. To train the model, image_alteration function was used to cut out pixels from the original image and return an image with missing parts that will be later used for training.
Used number of samples : 25k images
