Convolutional networks are typically used in image processing for object detection or recognition, classification etc. but not for generating images, as far as I know.