import cv2
import os
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
def detect_faces(img, draw_box=True):
# convert image to grayscale
grayscale_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# detect faces
faces = face_cascade.detectMultiScale(grayscale_img, scaleFactor=1.1,
minNeighbors=5,
minSize=(30, 30),
flags=cv2.CASCADE_SCALE_IMAGE)
face_box, face_coords = None, []
for (x, y, w, h) in faces:
if draw_box:
cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 0), 5)
face_box = img[y:y+h, x:x+w]
face_coords = [x,y,w,h]
return img, face_box, face_coords
if __name__ == "__main__":
files = os.listdir('sample_faces')
images = [file for file in files if 'jpg' in file]
for image in images:
img = cv2.imread('sample_faces/' + image)
detected_faces, _, _ = detect_faces(img)
cv2.imwrite('sample_faces/detected_faces/' + image, detected_faces)
