MTCNN+tensorflow测试人脸识别

据说MTCNN检测人脸十分靠谱，即使是侧脸也能检测得出，今天就查找资料做了个测试。有关MTCNN原理和训练等概不论述，只记录实现过程。

安装facenet，输入命令：

pip install facenet

安装完毕后找到detect_face.py文件目录，我的是win10系统，该文件在以下目录下：

C:\ProgramData\Anaconda3\Lib\site-packages\facenet\src\align

我就直接把align文件拷贝出来，和mtcnn.py文件放在一起，mtcnn.py文件内容为：

# coding=gbk
import tensorflow as tf
import align.detect_face
import cv2

minsize = 20 # minimum size of face
threshold = [ 0.6, 0.7, 0.7 ]  # three steps's threshold
factor = 0.709 # scale factor
gpu_memory_fraction=1.0
 
 
print('Creating networks and loading parameters')
 
with tf.Graph().as_default():
        gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=gpu_memory_fraction)
        sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options, log_device_placement=False))
        with sess.as_default():
            pnet, rnet, onet = align.detect_face.create_mtcnn(sess, None)

capture = cv2.VideoCapture(0)
while True:
    ret, frame = capture.read()
    if ret is True:
        bounding_boxes, _ = align.detect_face.detect_face(frame, minsize, pnet, rnet, onet, threshold, factor)
        for face_position in bounding_boxes:
           face_position=face_position.astype(int)
           cv2.rectangle(frame, (face_position[0], face_position[1]), (face_position[2], face_position[3]), (0, 255, 0), 2)  
        cv2.imshow("frame", frame)
        rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        c = cv2.waitKey(10)
        if c == 27:
            break
    else:
        break

在Windows下cmd进入命令行，找到mtcnn.py所在目录，输入命令：python mtcnn.py 

注意插上USB摄像头，win10安装了Python，TensorFlow，opencv-python

该程序在台式机上可实时监测，在树莓派3b+上实时效果并不理想