def gph_datasets_file_reorder_by_mapping(self, dir, gph_mapping_restore_file):
sys.stderr.write("Reordering gph files in %s by mapping"%gph_dir)
prefix = ''
choice = 1
instance = rename(dir, prefix, choice, gph_mapping_restore_file)
instance.run()
del instance #avoid the open file handlers
def datasets_reorder_by_sort(self, dir, organism, datasets_mapping_file):
sys.stderr.write("Reordering datasets in %s by sorting"%dir)
prefix = '%s_dataset'%organism
choice = 4
instance = rename(dir, prefix, choice, datasets_mapping_file)
instance.run()
del instance
def detect(self):
if not os.path.isfile(self.cascade_file):
raise RuntimeError("%s: not found" % self.cascade_file)
# rename files
if self.name == 1:
rename(self.src)
# Create face classifier
cascade = cv2.CascadeClassifier(self.cascade_file)
# get files
files = [
n for x in os.walk(self.src)
for n in glob(os.path.join(x[0], '*.*'))
]
# model loop
for image_file in files:
# create target path
target_dir = "/".join(image_file.strip("/").split('/')[1:-1])
target_dir = os.path.join(self.dst, target_dir) + "/"
# create target dir if doesn't exists
if not os.path.exists(target_dir):
os.makedirs(target_dir)
image = cv2.imread(image_file)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
gray = cv2.equalizeHist(gray)
faces = cascade.detectMultiScale(
gray,
# model options
scaleFactor=1.1,
minNeighbors=5,
minSize=(120, 120))
# crop, resize, and save image
WIDTH = 96
HEIGHT = 96
for (x, y, w, h) in faces:
crop_img = image[y:y + h, x:x + w]
crop_img = cv2.resize(crop_img, (WIDTH, HEIGHT),
interpolation=cv2.INTER_AREA)
filename = os.path.basename(image_file).split('.')[0]
cv2.imwrite(os.path.join(target_dir, filename + ".jpg"),
crop_img)
print("%s: has been cropped" % image_file)