目录

1. load_image_file 加载图像

import face_recognition

image = face_recognition.load_image_file("image.jpeg")   # 读取图像
print(image.shape)   # (3073, 2048, 3)
print(type(image))   # <class 'numpy.ndarray'>

2. face_locations 定位图中所有人脸

函数作用：识别出图像中所有人脸，并以列表的形式返回人脸的位置信息

函数输出：输出一个列表，形如： [(411, 956, 540, 827), (283, 605, 390, 497)]

import face_recognition
import cv2

image = face_recognition.load_image_file("image.jpeg")   # 读取图像
face_locations = face_recognition.face_locations(image)    # 获取人脸位置
print(face_locations)  

输出：

[(990, 1347, 1311, 1026), (491, 1133, 812, 812), (1895, 1152, 2163, 884)]

检测到的第一个人脸位置坐标(990, 1347, 1311, 1026)，对应如下

（1）在原图像中框出人脸

import face_recognition
import cv2

image = face_recognition.load_image_file("image.jpeg")   # 读取图像
face_locations = face_recognition.face_locations(image)    # 获取人脸位置
print(face_locations)

# ---------------------------------------------------
img = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
for p in face_locations:
    print(p)
    face_detect = cv2.rectangle(img, (p[3], p[0]), (p[1], p[2]), (0, 0, 255), 5)

cv2.imshow('face_detect'.format(), face_detect)
cv2.waitKey()
cv2.destroyAllWindows()

（2）剪切出人脸

import face_recognition
import cv2

image = face_recognition.load_image_file("image.jpeg")   # 读取图像
face_locations = face_recognition.face_locations(image)    # 获取人脸位置
print(face_locations)  # [(411, 956, 540, 827), (283, 605, 390, 497)]

img = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)

for i, p in enumerate(face_locations):
    face_detect = img[p[0]:p[2], p[3]:p[1]]
    cv2.imshow('face_detect_{}'.format(i), face_detect)
    
cv2.waitKey()
cv2.destroyAllWindows()

3. face_landmarks 识别人脸关键点

import face_recognition
import cv2
import numpy as np

image = face_recognition.load_image_file("image.jpeg")   # 读取图像
face_landmarks_list = face_recognition.face_landmarks(image)    # 获取人脸位置

img = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
for face_landmarks in face_landmarks_list:
    facial_features = [
        'left_eye',
        'right_eye',
        'nose_tip',
        'top_lip',
        'bottom_lip'
    ]


    # 在图像中画出每个人的眼睛， 鼻子， 嘴巴！
    for facial_feature in facial_features:
        points = np.array(face_landmarks[facial_feature])
        cv2.polylines(img, [points], False, (0, 0, 255), 2)

cv2.imshow('facial_feature', img)
cv2.waitKey()
cv2.destroyAllWindows()