OpenCV笔记_18(2)基于dlib进行人脸关键点检测(摄像头 )

18(2)基于dlib进行人脸关键点检测(摄像头 )

Tommy老师视频链接

凯哥英语视频

# 1.导入库
import cv2
import dlib

# 2.打开摄像头
capture = cv2.VideoCapture(0)

# 3.调用人脸检测器
detector = dlib.get_frontal_face_detector()

# 4.加载预测关键点模型(68个关键点)
# http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')

while True:
    # 5.读取视频流
    ret, frame = capture.read()
    # 6.灰度转转
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    # 7.人脸检测
    faces = detector(gray, 1)  # 0 原始大小 1 放大一倍
    # 8.循环,遍历每一张人脸,给人脸绘制矩形框和关键点
    for face in faces:
        # 8.1.绘制矩形框
        cv2.rectangle(frame, (face.left(), face.top()), (face.right(), face.bottom()), (0, 255, 0), 5)
        # 8.2.检测到关键点
        shape = predictor(gray, face)
        for pt in shape.parts():
            # 8.3.获取关键点坐标
            pt_position = (pt.x, pt.y)
            # 8.4.显示/绘制关键点
            cv2.circle(frame, pt_position, 2, (0, 0, 255), -1)
    # 9.退出判定
    if cv2.waitKey(1) & 0xff == ord('q'):
        break

    # 12.显示整个效果图
    cv2.imshow('face detection landamark', frame)

capture.release()
cv2.destroyAllWindows()

别的也没啥说的

ok，那就这样吧~

欢迎各位大佬留言吐槽,也可以深入交流~