def default_loader(path):
from torchvision import get_image_backend
if get_image_backend() == 'PIL':
return pil_loader(path)
else:
logging.info("Something went wrong with the default_loader in image_folder_segmentation.py")
sys.exit(-1)
def convert_gt(img_path):
img = np.array(pil_loader(img_path))
out_img = np.zeros(img.shape, dtype=np.uint8)
out_img[:, :, 2] = 1 # set everything to background in blue channel
out_img[:, :,
2][img[:, :, 2] != 0] = 2 # set glands to 2 in blue channel
out = Image.fromarray(out_img)
out.save(img_path)
def _load_test_images_and_vars(self):
"""
Inits the variables responsible of tracking which crop should be taken next, the current images and the like.
This should be run every time a new page gets loaded for the test-set
"""
# Load image
self.current_data_img = pil_loader(self.img_paths[self.next_image_index][0])
self.current_gt_img = pil_loader(self.img_paths[self.next_image_index][1])
# Initialize the sliding window indices
self.total_num_horiz_crops = self.num_horiz_crops[self.next_image_index]
self.current_horiz_crop = 0
self.total_num_vert_crops = self.num_vert_crops[self.next_image_index]
self.current_vert_crop = 0
self.total_crops_current_img = self.total_num_vert_crops * self.total_num_horiz_crops
# Update pointer to next image
self.next_image_index += 1
def layout_analysis_output(args):
ground_truth_folder = args.ground_truth_folder
network_output_folder = args.network_output_folder
output_folder = os.path.join(args.output_folder,
'layout_analysis_evaluation')
gt_img_paths = get_img_paths(ground_truth_folder)
segm_img_paths = get_img_paths(network_output_folder)
if not output_folder:
output_folder = os.path.join(os.path.dirname(segm_img_paths),
'layout_analysis_evaluation')
for gt, segm in zip(gt_img_paths, segm_img_paths):
gt_image = np.array(pil_loader(gt))
segm_image = np.array(pil_loader(segm))
generate_layout_analysis_output(output_folder, gt_image, segm_image,
os.path.basename(gt))
# create a legend and save it
make_colour_legend_image(
os.path.join(output_folder, "layout_analysis_eval_legend"),
CLASS_COLOUR_ENCODINGS)
def _load_image(self, input_folder):
"""Load the image from the file system"""
if not os.path.exists(input_folder):
raise FileNotFoundError(f"Could not find file {input_folder}")
img = pil_loader(input_folder)
return img