I'm already quite interested in resizing wide images to square images for Convolutional Neural Networks training