create_bckgnd.py

import numpy as np

#######################
# Concatenates the front-light, back-light (and groundtruth) images to a 6-/7-channel image
#######################

### HORIZONTAL YARNS = 1
### VERTICAL YARNS = 0

#def transform_patch(p):
#    p1 = np.swapaxes(p, 0, 2)
#    return p1

#######################

def create_7s_images(front, back, groundtruth=None):
    if groundtruth is None:
        res = np.concatenate((front, back), axis=2)
    else:
        groundtruth = np.expand_dims(np.uint8(np.round(groundtruth[:, :, 0]/128)), axis=2)
        res = np.concatenate((front, back, groundtruth), axis=2)
    return res