Simple face detection using mediapipe and opencv

# Face detection 
# opencv
import cv2
# mediapipe ai tool kit 
import mediapipe as mp
#  Progress bar library 
from tqdm import tqdm
import time
import matplotlib.pyplot as plt



#  Define a visual image function 
def look_img(img):
    img_RGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    plt.imshow
    plt.show()


#  Import BlazeFace Model 
mp_face_detection = mp.solutions.face_detection
model = mp_face_detection.FaceDetection(
    min_detection_confidence=0.5,  #  Confidence threshold , Filter out prediction boxes that are less than confidence 
    model_selection=1,  #  Choose a model ,0  It is applicable when the face is close to the camera (2 Minet ),1  Suitable for far away (5 Within meters )
)
#  Import visualization functions and visualization styles 
mp_drawing=mp.solutions.drawing_utils
#  Key style 
keypoint_style=mp_drawing.DrawingSpec(thickness=5,circle_radius=3,color=(0,255,0))
#  Face prediction box style 
bbox_style=mp_drawing.DrawingSpec(thickness=5,circle_radius=3,color=(255,0,0))



#  Read image 
img = cv2.imread('images/img.png')
# BGR turn RGB
img_RGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
#  take RGB Input the prediction results of the model 
results=model.process(img_RGB)

#  Visualize face frames and face keys 
annotated_image=img.copy()
for detection in results.detections:
    mp_drawing.draw_detection(annotated_image,
                              detection,
                              keypoint_drawing_spec=keypoint_style,
                              bbox_drawing_spec=bbox_style)
look_img(annotated_image)
cv2.imwrite('Test.jpg',annotated_image)

material

Treatment effect