“Wykrywanie siatkówki” Kod odpowiedzi

Wykrywanie siatkówki

#pip3 install opencv-python
import cv2 
from retinaface import RetinaFace

# init with normal accuracy option
detector = RetinaFace(quality="normal")

# same with cv2.imread,cv2.cvtColor 
rgb_image = detector.read("data/hian.jpg")

faces = detector.predict(rgb_image)
# faces is list of face dictionary
# each face dictionary contains x1 y1 x2 y2 left_eye right_eye nose left_lip right_lip
# faces=[{"x1":20,"y1":32, ... }, ...]

result_img = detector.draw(rgb_image,faces)

# save ([...,::-1] : rgb -> bgr )
cv2.imwrite("data/result_img.jpg",result_img[...,::-1])

# show using cv2
# cv2.imshow("result",result_img[...,::-1)
# cv2.waitKey()