-
Notifications
You must be signed in to change notification settings - Fork 0
/
norm_mask.py
51 lines (44 loc) · 1.48 KB
/
norm_mask.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
import os
import argparse
import numpy as np
import cv2
MAP = {(255, 255, 255): 1}
OTHERS = 0
# Credit: @heaversm at GitHub
# https:/heaversm/deeplab-training/blob/master/models/research/deeplab/datasets/convert_rgb_to_index.py
def norm_mask(mask):
normed = np.full((mask.shape[0], mask.shape[1]), OTHERS, dtype=np.uint8)
for c, i in MAP.items():
m = np.all(mask == np.array(c).reshape(1, 1, 3), axis=2)
normed[m] = i
return normed
def get_files(path):
files = os.listdir(path)
files = [f for f in files if f.endswith('.png')]
files.sort()
return files
def read_mask(mask_file):
mask = cv2.imread(mask_file)
mask = cv2.cvtColor(mask, cv2.COLOR_BGR2RGB)
return mask
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("-m", "--masks" , help="Path to masks", required=True)
parser.add_argument("-n", "--norm-masks", help="Path to normalized masks")
args = parser.parse_args()
print(args)
return args
def main():
args = parse_args()
masks_path = args.masks
norm_masks_path = args.norm_masks
if not os.path.exists(norm_masks_path):
os.mkdir(norm_masks_path)
masks = get_files(masks_path)
for mask_file in masks:
print(f'[INFO] Processing {mask_file}...')
mask = read_mask(os.path.join(masks_path, mask_file))
normed = norm_mask(mask)
cv2.imwrite(os.path.join(norm_masks_path, mask_file), normed)
if __name__ == '__main__':
main()