Remake me this model remove the text from the image and only generate the model and use the front of the image to guess the back and make sure there is no gaps in the model only the middle