使用MTACNN进行人脸检测

github地址

https://github.com/ipazc/mtcnn

本机环境

python 3.7
Keras 2.2.0
opencv-python 4.5.1.48
tensorflow 1.13.1

检测函数及实例

import cv2
from mtcnn import MTCNN

detector = MTCNN()

def detec_face(imgFileName):
    image = cv2.cvtColor(cv2.imread(imgFileName), cv2.COLOR_BGR2RGB)
    result = detector.detect_faces(image)

    if len(result) == 0:
        print("图片 "+imgFileName+" 没有检测到人脸。")
        print("---------------------------")
    else:
        count = len(result)
        for item in result:
            bounding_box = item["box"]
            cv2.rectangle(image,
                          (bounding_box[0], bounding_box[1]),
                          (bounding_box[0] + bounding_box[2], bounding_box[1] + bounding_box[3]),
                          (0, 155, 255),
                          2)
            cv2.imwrite(imgFileName.split(".")[0]+"_result."+imgFileName.split(".")[1], cv2.cvtColor(image, cv2.COLOR_RGB2BGR))
        print("图片 "+imgFileName+" 检测到{}张人脸，如下：".format(count))
        for i in range(count):
            print("人脸{}：{}".format(i+1,result[i]))
        print("---------------------------")

detec_face("image/single_face.jpg")
detec_face("image/more_face.jpg")
detec_face("image/no_face.jpg")

运行结果

程序输出：

左侧原图，右侧运行结果
no_face.jpg为不含人脸的风景图片，不做列出

标签：box,MTACNN,检测,image,cv2,face,人脸,result,bounding
来源： https://blog.csdn.net/weixin_44013961/article/details/115260925