def calib_input(iter):
images = []
line = open(calib_image_list).readlines()
#print(line)
for index in range(0, calib_batch_size):
curline = line[iter * calib_batch_size + index]
#print("iter= ", iter, "index= ", index, "sum= ", int(iter*calib_batch_size + index), "curline= ", curline)
calib_image_name = curline.strip()
image_path = os.path.join(calib_image_dir, calib_image_name)
image2 = cnn.NormalizeImageArr(image_path)
#image2 = image2.reshape((image2.shape[0], image2.shape[1], 3))
images.append(image2)
return {"input_1": images}
else:
model = unet.UNET_v3(N_CLASSES, HEIGHT, WIDTH)
######################################################################
# prepare testing and validation data
######################################################################
# load validation images
valid_images = os.listdir(dir_valid_img)
valid_images.sort()
valid_segmentations = os.listdir(dir_valid_seg)
valid_segmentations.sort()
X_valid = []
Y_valid = []
for im, seg in zip(valid_images, valid_segmentations):
X_valid.append(cnn.NormalizeImageArr(os.path.join(dir_valid_img, im)))
Y_valid.append(
cnn.LoadSegmentationArr(os.path.join(dir_valid_seg, seg), N_CLASSES,
WIDTH, HEIGHT))
X_valid, Y_valid = np.array(X_valid), np.array(Y_valid)
print("\n")
print("validation data (X) (Y) shapes:", X_valid.shape, Y_valid.shape)
# load testing images
test_images = os.listdir(dir_test_img)
test_images.sort()
test_segmentations = os.listdir(dir_test_seg)
test_segmentations.sort()
X_test = []
Y_test = []
for im, seg in zip(test_images, test_segmentations):