人脸检测器

from mtcnn.mtcnn import MTCNN
import cv2
img=cv2.imread('../data/test/2.jpg')
detector=MTCNN()
box=detector.detect_faces(img)
faces=[]
for i in range(len(box)):
    faces.append(box[i]['box'])
for (x, y, w, h) in faces:
    print(x)

mtcnn速度慢

cascade = cv2.CascadeClassifier('../data/params/haarcascade_frontalface_default.xml')  # 检测人脸
faces = cascade.detectMultiScale(frame_gray, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30),
                                                flags=cv2.CASCADE_SCALE_IMAGE)

不精准

dlib

def detection(img):
    """
    人脸检测
    :param img:
    :return:
    """
    detector = dlib.get_frontal_face_detector()
    dets = detector(img, 1)  # 使用detector进行人脸检测 dets为返回的结果
    print("Number of faces detected: {}".format(len(dets)))  # 打印识别到的人脸个数
    for index, face in enumerate(dets):
        # 在图片中标注人脸，并显示
        left = face.left()
        top = face.top()
        right = face.right()
        bottom = face.bottom()
        cv.rectangle(img, (left, top), (right, bottom), (0, 255, 0), 1)
    return dets

68个标注点定位

def predictor(img, dets):
    """
    特征点标定
    :param img:
    :param dets:
    :return:
    """
    # shape_predictor_68_face_landmarks.dat是进行人脸标定的模型，它是基于HOG特征的，这里是他所在的路径
    predictor_path = "../model/shape_predictor_68_face_landmarks.dat"
    predictor = dlib.shape_predictor(predictor_path)
    shape_list = []
    for index, face in enumerate(dets):
        shape = predictor(img, face)  # 寻找人脸的68个标定点
        # 遍历所有点，打印出其坐标，并用绿色的圈表示出来
        for _, pt in enumerate(shape.parts()):
            pt_pos = (pt.x, pt.y)
            cv.circle(img, pt_pos, 1, (0, 255, 0), 1)
        shape_list.append(shape)
    return shape_list