File size: 6,531 Bytes
54fa0c8 |
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 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 |
'''
MIT license https://opensource.org/licenses/MIT Copyright 2024 Infosys Ltd
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
'''
from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
from tensorflow.keras.preprocessing.image import img_to_array
from tensorflow.keras.models import load_model
from privacy.config.logger import CustomLogger
import numpy as np
import argparse
import cv2
import os
import argparse
import random
import os
from tqdm import tqdm
log= CustomLogger()
parser = argparse.ArgumentParser()
parser.add_argument('-d', '--data-dir', default='data/dataset_raw',
help="Directory with the raw dataset")
parser.add_argument('-o', '--output-dir', default='data/64x64_dataset',
help="Where to write the new data")
parser.add_argument('-s', '--size', type=int, default=64,
help="Where to write the new data")
parser.add_argument('-c', '--confidence', type=float, default=0.5,
help="Confidence threshold to detect face")
parser.add_argument('--face-model', type=str, default="face_detector",
help="path to face detector model directory")
def extract_face(filename, output_dir, net, size, confidence_threshold):
image = cv2.imread(filename)
if image is None:
return
filename_out = filename.split('/')[-1].split('.')[0]
(h, w) = image.shape[:2]
blob = cv2.dnn.blobFromImage(image, scalefactor=1.0, size=(128, 128), mean=(104.0, 177.0, 123.0))
net.setInput(blob)
detections = net.forward()
for i in range(0, detections.shape[2]):
confidence = detections[0, 0, i, 2]
if confidence > confidence_threshold:
box = detections[0, 0, i, 3:7] * np.array([w, h, w, h])
(startX, startY, endX, endY) = box.astype("int")
(startX, startY) = (max(0, startX), max(0, startY))
(endX, endY) = (min(w - 1, endX), min(h - 1, endY))
try:
frame = image[startY:endY, startX:endX]
frame = cv2.resize(frame, (size, size), interpolation=cv2.INTER_AREA)
if i > 0:
image_out = os.path.join(output_dir, '%s_%s.jpg' % (filename_out, i))
else:
image_out = os.path.join(output_dir, '%s.jpg' % filename_out)
cv2.imwrite(image_out, frame)
except Exception as e:
log.error(str(e))
log.error("Line No:"+str(e.__traceback__.tb_lineno))
def app():
args = parser.parse_args()
assert os.path.isdir(args.data_dir), "Couldn't find the dataset at {}".format(args.data_dir)
prototxtPath = os.path.join(args.face_model, "deploy.prototxt")
weightsPath = os.path.join(args.face_model, "res10_300x300_ssd_iter_140000.caffemodel")
net = cv2.dnn.readNet(prototxtPath, weightsPath)
# Define the data directories
train_mask_dir = os.path.join(args.data_dir, 'train/Mask')
train_non_mask_dir = os.path.join(args.data_dir, 'train/Non Mask')
os.makedirs(train_mask_dir, exist_ok=True)
os.makedirs(train_non_mask_dir, exist_ok=True)
test_mask_dir = os.path.join(args.data_dir, 'test/Mask')
test_non_mask_dir = os.path.join(args.data_dir, 'test/Non Mask')
os.makedirs(test_mask_dir, exist_ok=True)
os.makedirs(test_non_mask_dir, exist_ok=True)
# Get the filenames in each directory (train and test)
filenames_mask = os.listdir(train_mask_dir)
filenames_mask = [os.path.join(train_mask_dir, f) for f in filenames_mask if f.endswith('.jpg')]
filenames_non_mask = os.listdir(train_non_mask_dir)
filenames_non_mask = [os.path.join(train_non_mask_dir, f) for f in filenames_non_mask if f.endswith('.jpg')]
test_filenames_mask = os.listdir(test_mask_dir)
test_filenames_mask = [os.path.join(test_mask_dir, f) for f in test_filenames_mask if f.endswith('.jpg')]
test_filenames_non_mask = os.listdir(test_non_mask_dir)
test_filenames_non_mask = [os.path.join(test_non_mask_dir, f) for f in test_filenames_non_mask if f.endswith('.jpg')]
# Split the images into 80% train and 20% dev
# Make sure to always shuffle with a fixed seed so that the split is reproducible
filenames_mask.sort()
filenames_non_mask.sort()
random.shuffle(filenames_mask)
random.shuffle(filenames_non_mask)
split_mask = int(0.8 * len(filenames_mask))
train_filenames_mask = filenames_mask[:split_mask]
dev_filenames_mask = filenames_mask[split_mask:]
split_non_mask = int(0.8 * len(filenames_non_mask))
train_filenames_non_mask = filenames_non_mask[:split_non_mask]
dev_filenames_non_mask = filenames_non_mask[split_non_mask:]
filenames = {'train/Mask': train_filenames_mask,
'train/Non Mask': train_filenames_non_mask,
'test/Mask': test_filenames_mask,
'test/Non Mask': test_filenames_non_mask,
'validation/Mask': dev_filenames_mask,
'validation/Non Mask': dev_filenames_non_mask}
# Preprocess train, dev and test
for split in filenames.keys():
output_dir_split = os.path.join(args.output_dir, split)
os.makedirs(output_dir_split, exist_ok=True)
print("Processing {} data, saving preprocessed data to {}".format(split, output_dir_split))
for filename in tqdm(filenames[split]):
extract_face(filename, output_dir_split, net, args.size, args.confidence)
print("Done building dataset")
if __name__ == '__main__':
app()
|