在加载用于遮罩检测的keras模型时，使用mtcnn的面部检测模型会停止工作，并且一旦load_model行被注释，它就会开始工作吗？我已经检查了功能本身，那里没有问题。这是代码：
CUDA_VISIBLE_DEVICES = 0
parser = argparse.ArgumentParser()
parser.add_argument('--check', help='if its video or folder.', type=str, default=0)
args = parser.parse_args()
print(sys.argv)
log = pd.DataFrame(columns=["name", "date and time"])
cuda = clear_cuda_memory()
if args.check == 'video':
#
MINSIZE = 20
THRESHOLD = [0.6, 0.7, 0.7]
FACTOR = 0.709
IMAGE_SIZE = 182
INPUT_IMAGE_SIZE = 160
CLASSIFIER_PATH = '/home/ubuntu/FacRecog/attendance_system_face_recognition/facenet-master/face-recognition-master/dataset/Friends/innovo_528_GPU'
mode_for_names={}
VIDEO_PATH = 'C:\\Users\\Lenovo\\Downloads\\Data_science\\it.mkv'
FACENET_MODEL_PATH = '/home/ubuntu/FacRecog/attendance_system_face_recognition/facenet-master/face-recognition-master/dataset/Friends/Models/new/20180408-102900.pb'
unknown_count = 0
person_count = 0
workbook = xlsxwriter.Workbook('log.xlsx')
worksheet = workbook.add_worksheet()
worksheet.write('A1', 'name')
worksheet.write('B1', 'date and time')
# Load The Custom Classifier
with open(CLASSIFIER_PATH, 'rb') as file:
model, class_names = pickle.load(file)
print("Custom Classifier, Successfully loaded")
with tf.Graph().as_default():
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.2)
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options, log_device_placement=False))
with sess.as_default():
# Load the model
print('Loading feature extraction model')
# facenet.load_model(FACENET_MODEL_PATH)
facenet.load_model(FACENET_MODEL_PATH)# load model for face recognition
# Get input and output tensors
images_placeholder = tf.get_default_graph().get_tensor_by_name("input:0")
embeddings = tf.get_default_graph().get_tensor_by_name("embeddings:0")
phase_train_placeholder = tf.get_default_graph().get_tensor_by_name("phase_train:0")
embedding_size = embeddings.get_shape()[1]
pnet, rnet, onet = align.detect_face.create_mtcnn(sess, "./src/align")# used to find the onet pent rnet for mtcnn
people_detected = set()
person_detected = collections.Counter()
phone = 'https://192.168.18.64:8080/video'
# ip_cam='admin:123@192.168.18.157/media/?action=stream'
computer = 0
cap = cv2.VideoCapture(phone)
# fvs = FileVideoStream(phone).start()
excel = 1
check_for_mode=0
check_for_unknown=0
max_one = -9999
# face_clsfr = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
mask_check = 0
size = 2
webcam = cv2.VideoCapture('https://192.168.18.64:8080/video') # Use camera 0
label = 5
mask_count = 0
# We load the xml file
classifier = cv2.CascadeClassifier('/home/ubuntu/opencv-3.3.0/data/haarcascades/haarcascade_frontalface_default.xml')
model_mask = load_model("model8-008.model")#the main issue exits in this line as as soon as it is commented the code works
labels_dict = {0: 'without mask', 1: 'mask'}
color_dict = {0: (0, 0, 255), 1: (0, 255, 0)}
while True:
try:
(rval, im) = webcam.read()
bounding_boxes, _ = align.detect_face.detect_face(im, MINSIZE, pnet, rnet, onet, THRESHOLD,
FACTOR)# used for mtcnn face detection
faces_found = bounding_boxes.shape[0]
# gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# faces = face_clsfr.detectMultiScale(frame, 1.3, 5)
max_check = []
try:
if faces_found > 0:
det = bounding_boxes[:, 0:4]
bb = np.zeros((faces_found, 4), dtype=np.int32)
face = 0
cv2.waitKey(1)
for i in range(faces_found):
bb[i][0] = det[i][0]
bb[i][1] = det[i][1]
bb[i][2] = det[i][2]
bb[i][3] = det[i][3]
x = np.sqrt((bb[i][2] - bb[i][0]) ** 2 + (bb[i][3] - bb[i][1]) ** 2)
max_check.append(x)
value_max = np.max(max_check)
index_max = max_check.index(value_max)
face = index_max
cropped = im[bb[face][1]:bb[face][3], bb[face][0]:bb[face][2], :]
resized=cv2.resize(cropped,(150,150))
normalized=resized/255.0
reshaped=np.reshape(normalized,(1,150,150,3))
reshaped = np.vstack([reshaped])
resized1=cv2.resize(cropped,(150,150))
normalized1=resized1/255.0
reshaped1=np.reshape(normalized1,(1,150,150,3))
reshaped1 = np.vstack([reshaped1])
result=model_mask.predict(reshaped1)
print(result)
label=np.argmax(result,axis=1)[0]
#
# cv2.rectangle(im,(x,y),(x+w,y+h),color_dict[label],2)
# cv2.rectangle(im,(x,y-40),(x+w,y),color_dict[label],-1)
# cv2.putText(im, labels_dict[label], (x, y-10),cv2.FONT_HERSHEY_SIMPLEX,0.8,(255,255,255),2)
except Exception as e:
print(e)
except Exception as e:
print(e)
如果代码的显示或结构存在任何问题，请告知。