def __init__(self):
""" initialization
mtcnn_model：Used to extract face frame, output 5 key points (eyes, mouth, nose tip)
facenet_model：Face recognition model, output 128 individual face feature points
"""
self.mtcnn_model = mtcnn.MTCNN()
self.threshold = [0.5, 0.6, 0.8]
self.facenet_model = InceptionResNetV1()
model_path = 'model/facenet_keras.h5'
self.facenet_model.load_weights(model_path)
self.known_face_encodings = []
self.known_face_names = []
# Import face library face_date.pkl
if os.path.exists('model/face_date.pkl'):
with open('model/face_date.pkl', 'rb') as fr:
try:
data = pickle.load(fr)
self.known_face_encodings = data[0]
self.known_face_names = data[1]
except EOFError:
print("face_date.pkl文件为空")
else:
with open("model/face_date.pkl", mode='w', encoding='utf-8') as ff:
pass
def face_extract():
print('Face The Camera')
camera = cv2.VideoCapture(0)
time.sleep(1)
return_value, image = camera.read()
del camera
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
pixels = np.asarray(image)
detector = mtcnn.MTCNN()
results = detector.detect_faces(pixels)
# extract the bounding box from the first face
x1, y1, width, height = results[0]['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
# extract the face
face = pixels[y1:y2, x1:x2]
image = Image.fromarray(face)
image = image.resize((160, 160))
face_array = np.asarray(image)
face_array = np.expand_dims(face_array, axis=0)
return face_array
def read_frames(self):
# start the file video stream thread and allow the buffer to
# start to fill
print("[INFO] starting video file thread...")
args = self.argument_parser()
output_path = self.cwd / Path(args.output_dir)
output_path.mkdir(parents=True, exist_ok=True)
fvs = FileVideoStream(args.video).start()
time.sleep(1.0)
# start the FPS timer
fps = FPS().start()
detector = mtcnn.MTCNN()
fno = 0
frames = VideoMeta(args.video).fps()
#print(meta)
# loop over frames from the video file stream
while fvs.more():
fno += 1
print(fno)
# grab the frame from the threaded video file stream, resize
# it, and convert it to grayscale (while still retaining 3
# channels)
frame = fvs.read()
if (fno % frames != 0):
continue
try:
frame = imutils.resize(frame, width=450)
except:
break
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# frame = np.dstack([frame, frame, frame])
if args.detect == "haarcascade":
results = self.face_cascade.detectMultiScale(frame, 1.3, 5)
if len(results) != 0:
self.crop(frame, results[0],
str(output_path / Path(str(fno) + ".jpg")))
else:
results = detector.detect_faces(frame)
# display the size of the queue on the frame
# cv2.putText(frame, "Queue Size: {}".format(fvs.Q.qsize()),
# (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
# show the frame and update the FPS counter
cv2.imshow("Frame", frame)
cv2.waitKey(1)
fps.update()
# stop the timer and display FPS information
fps.stop()
print("[INFO] elasped time: {:.2f}".format(fps.elapsed()))
print("[INFO] approx. FPS: {:.2f}".format(fps.fps()))
# do a bit of cleanup
cv2.destroyAllWindows()
fvs.stop()
def get_frame(self):
conn = sqlite3.connect('database.db')
c = conn.cursor()
fname = "recognizer/trainingData.yml"
if not os.path.isfile(fname):
print("\nPlease train the data first\n")
return None
# Model For Facial Expression Recognition
model = FacialExpressionModel("model.json", "model_weights.h5")
# Model For Face Recognizer
face_detector = mtcnn.MTCNN()
recognizer = cv2.face.LBPHFaceRecognizer_create()
recognizer.read(fname)
_, img = self.video.read()
faces = face_detector.detect_faces(img)
for res in faces:
x, y, w, h = res["box"]
x, y = abs(x), abs(y)
x1, y1 = x+w, y+h
image = img[y:y1, x:x1]
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# Recognizing Face
ids, conf = recognizer.predict(image)
c.execute("select roll_no from users where id = (?);", (ids,))
result = c.fetchall()
try:
roll_no = result[0][0]
except:
roll_no = 'Error'
if conf > 50:
roll_no = "No Match"
image2 = cv2.resize(image, (48, 48))
# Predicting Expression
pred = model.predict_emotion(image2[np.newaxis, :, :, np.newaxis])
msg = pred + " " + roll_no
# Mark the Expression if Face is detected
if roll_no != "Error" and roll_no != "No Match" :
marked = count(roll_no, pred)
if(marked):
Attendance(roll_no)
msg = "MARKED"
cv2.putText(img, msg, (x, y), cv2.FONT_HERSHEY_SIMPLEX,
1, (255, 255, 0), 2)
cv2.rectangle(img, (x, y), (x+w, y+h), (255, 0, 0), 2)
_, jpeg = cv2.imencode('.jpg', img)
return jpeg.tobytes()
def read_frames(self):
# start the file video stream thread and allow the buffer to
# start to fill
print("[INFO] starting video file thread...")
args = self.argument_parser()
output_path = self.cwd / Path(args.output_dir)
output_path.mkdir(parents=True, exist_ok=True)
fvs = FileVideoStream(args.video).start()
time.sleep(1.0)
# start the FPS timer
fps = FPS().start()
detector = mtcnn.MTCNN()
fno = 0
#frames = VideoMeta(args.video).fps()
#print(meta)
e = TfPoseEstimator(get_graph_path("mobilenet_thin"),
target_size=(432, 368))
# loop over frames from the video file stream
while fvs.more():
fno += 1
print(fno)
# grab the frame from the threaded video file stream, resize
# it, and convert it to grayscale (while still retaining 3
# channels)
frame = fvs.read()
#if (fno % frames != 0):
# continue
try:
frame = imutils.resize(frame, width=432, height=368)
except:
break
#frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#frame = np.dstack([frame, frame, frame])
humans = e.inference(frame)
image = TfPoseEstimator.draw_humans(frame, humans, imgcopy=False)
print("humans:", humans)
# display the size of the queue on the frame
# cv2.putText(frame, "Queue Size: {}".format(fvs.Q.qsize()),
# (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
# show the frame and update the FPS counter
cv2.imshow("Frame", image)
cv2.waitKey(1)
fps.update()
# stop the timer and display FPS information
fps.stop()
print("[INFO] elasped time: {:.2f}".format(fps.elapsed()))
print("[INFO] approx. FPS: {:.2f}".format(fps.fps()))
# do a bit of cleanup
cv2.destroyAllWindows()
fvs.stop()
def Entry(uname, your_roll_no, branch):
# Connecting database
conn = sqlite3.connect('database.db')
if not os.path.exists('dataset'):
os.makedirs('dataset')
c = conn.cursor()
face_detector = mtcnn.MTCNN()
cap = cv2.VideoCapture(0)
new_folder = str(your_roll_no) + "-" + str(uname) + "-" + str(branch)
if not os.path.exists("dataset/" + new_folder):
os.mkdir("dataset/" + new_folder)
else:
return "Student already exists!!!"
c.execute('INSERT INTO users (roll_no) VALUES (?)', (your_roll_no, ))
uid = c.lastrowid
sampleNum = 0
while True:
ret, img = cap.read()
# Detecting faces from live Camera and cropping the faces
faces = face_detector.detect_faces(img)
for res in faces:
x, y, w, h = res['box']
x, y = abs(x), abs(y)
x1, y1 = x + w, y + h
sampleNum = sampleNum + 1
image = img[y:y1, x:x1]
image = cv2.resize(image, (256, 320))
# Saving the cropped faces with ID
cv2.imwrite(
"dataset/" + new_folder + "/Student." + str(uid) + "." +
str(sampleNum) + ".jpg", image)
register(uname, your_roll_no, branch)
cv2.rectangle(img, (x, y), (x1, y1), (255, 0, 0), 2)
cv2.waitKey(20)
cv2.waitKey(1)
if sampleNum >= 15:
break
cap.release()
conn.commit()
conn.close()
cv2.destroyAllWindows()
# Training the Dataset
TrainDataset()
return "Student Data Registered!!!"
def detect_face(img):
pixels = np.asarray(img)
detector = mtcnn.MTCNN()
faces = detector.detect_faces(pixels)
try:
if(len(faces)!=1):
raise DetectionError
except DetectionError:
print("Either no or multiple faces were found in the captured image.")
pass
x1, y1, width, height = faces[0]['box']
x2 = x1 + width
y2 = y1 + height
face = pixels[y1:y2, x1:x2]
return face
def preprocessing(directory_path, save_path, required_size=(225, 225)):
detector = dlib.get_frontal_face_detector()
detecto_mtcnn = mtcnn.MTCNN()
train_paths = glob(fr"{directory_path}\*")
if not os.path.exists(f"{save_path}"):
os.mkdir(fr"{save_path}")
else:
pass
for path in tqdm(train_paths):
name = path.split("\\")[-1]
images = glob(f"{path}\\*")
for image_path in images:
try:
temp_path = image_path.split("\\")[-1]
image = cv2.imread(image_path)
# image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
face_recs = detector(image, 1)
face = None
if len(face_recs) == 1:
x = face_recs[0].left()
y = face_recs[0].top()
w = face_recs[0].right() - x
h = face_recs[0].bottom() - y
face = cv2.resize(image[y:y + h, x:x + w], required_size, interpolation=cv2.INTER_LINEAR)
else:
print(fr"{name}")
faces = detecto_mtcnn.detect_faces(image)
for face in faces:
x, y, width, height = face['box']
face = cv2.resize(image[y:y + height, x:x + width], required_size,
interpolation=cv2.INTER_LINEAR)
if not os.path.exists(fr"{save_path}\{name}"):
os.mkdir(fr"{save_path}\{name}")
cv2.imwrite(fr"{save_path}\{name}\{temp_path}", face)
else:
cv2.imwrite(fr"{save_path}\{name}\{temp_path}", face)
except Exception as e:
print(e)
exc_info = sys.exc_info()
traceback.print_exception(*exc_info)
del exc_info
pass
def main():
ch = 'y'
detector = mtcnn.MTCNN()
while(ch=='y' or ch=='Y'):
try:
cap = cv2.VideoCapture(0)
except:
sys.exit('An error occured while initializing the camera.')
i=0
name = input('Enter name : ')
roll = input('Enter roll no. : ')
year = input('Enter graduation year : ')
branch = input('Enter branch : ')
isImageDirPresent, path = imagePath(roll)
if isImageDirPresent=='end':
continue
print('Directory created successfully.')
print('Alright. Good to go. Say cheese...')
# To capture an image, press 'c'. To end capturing, press (or long press) 'q'.
while(True):
try:
mydir = path
ret, frame = cap.read()
frame = cv2.flip(frame, 1)
cv2.imshow('Capturing', frame)
if cv2.waitKey(1) & 0xff==ord('c'):
face = detectFace(frame)
filename = os.path.join(path, str(i) + '.jpg')
# print(filename)
cv2.imwrite(filename, frame)
i = i+1
elif cv2.waitKey(1) & 0xff==ord('q'):
break
except:
print("Something went wrong")
cap.release()
cv2.destroyAllWindows()
print('Saved {} images of {}.'.format(i, name))
print('')
ch = input('Run again ? (y/n) : ')
print('*******end*******')
def face_extract(file, size):
image = cv2.imread(file)
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
pixels = np.asarray(image)
detector = mtcnn.MTCNN()
results = detector.detect_faces(pixels)
# extract the bounding box from the first face
x1, y1, width, height = results[0]['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
# extract the face
face = pixels[y1:y2, x1:x2]
image = Image.fromarray(face)
image = image.resize(size)
face_array = np.asarray(image)
return face_array
def collect_data(img):
detector = mtcnn.MTCNN()
image_size = detector.image_size
input_image_size = detector.input_image_size
bounding_boxes, img = detector.run_mtcnn(img)
nrof_faces = bounding_boxes.shape[0]
if nrof_faces > 0:
det = bounding_boxes[:, 0:4]
img_size = np.asarray(img.shape)[0:2]
if nrof_faces > 1:
bounding_box_size = (det[:, 2] - det[:, 0]) * (det[:, 3] -
det[:, 1])
img_center = img_size / 2
offsets = np.vstack([(det[:, 0] + det[:, 2]) / 2 - img_center[1],
(det[:, 1] + det[:, 3]) / 2 - img_center[0]])
offset_dist_squared = np.sum(np.power(offsets, 2.0), 0)
index = np.argmax(bounding_box_size - offset_dist_squared * 2.0)
det = det[index, :]
det = np.squeeze(det)
bb_temp = np.zeros(4, dtype=np.int32)
bb_temp[0] = det[0]
bb_temp[1] = det[1]
bb_temp[2] = det[2]
bb_temp[3] = det[3]
cropped_temp = img[bb_temp[1]:bb_temp[3], bb_temp[0]:bb_temp[2], :]
cropped_temp = facenet.flip(cropped_temp, False)
scaled_temp = misc.imresize(cropped_temp, (image_size, image_size),
interp='bilinear')
scaled_temp = cv2.resize(scaled_temp,
(input_image_size, input_image_size),
interpolation=cv2.INTER_CUBIC)
scaled_temp = facenet.prewhiten(scaled_temp)
scaled_reshape = scaled_temp.reshape(-1, input_image_size,
input_image_size, 3)
return scaled_reshape
def extract_face(filename, required_size=(160, 160)):
# print(filename)
image = Image.open(filename)
image = image.convert('RGB')
pixels = np.asarray(image)
detector = mtcnn.MTCNN()
faces = detector.detect_faces(pixels)
if len(faces)==0:
#print('No faces were detected in the image {}'.format(filename))
return 'end', 'end'
x1, y1, width, height = faces[0]['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
face = pixels[y1:y2, x1:x2]
image = Image.fromarray(face)
image = image.resize(required_size)
image.save(filename)
face_array = np.asarray(image)
return face_array, 'true'
def recognizeEmotion(model):
pixels = cv2.cvtColor(get_new_img_webcam(), cv2.COLOR_BGR2RGB)
detector = mtcnn.MTCNN()
faces = detector.detect_faces(pixels)
print(faces)
x, y, width, height = faces[0]['box']
face = pixels[y:y + height, x:x + width]
fig, axs = plt.subplots(1, 2, figsize=(12, 5))
axs[0].imshow(pixels)
axs[1].imshow(face)
plt.show()
face = Image.fromarray(face, 'RGB')
face = transforms.Resize((224, 224))(face)
face = transforms.ToTensor()(face).unsqueeze(1)
# https://stackoverflow.com/questions/56789038/runtimeerror-given-groups-1-weight-of-size-64-3-3-3-expected-input4-50
face = face.permute(1, 0, 2, 3)
output = model(face)
pred = output.argmax(dim=1, keepdim=True)
class_ = getClassById(pred)
return class_
def save_img():
key = cv2.waitKey()
webcam = cv2.VideoCapture(0)
while True:
try:
check, frame = webcam.read()
cv2.imshow("capturing", frame)
key = cv2.waitKey(1)
if key == ord('s'):
cv2.imshow("captured image", frame)
pixels = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
detector = mtcnn.MTCNN()
results = detector.detect_faces(pixels)
x1, y1, width, height = results[0]['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
#face=pixels[y1:y2,x1:x2]
face = pixels[y1 - 15:y2 + 15, x1 - 15:x2 + 15]
image = Image.fromarray(face)
image = image.resize((96, 96))
face = np.array(image)
opencvImage = cv2.cvtColor(face, cv2.COLOR_RGB2BGR)
n = input("Enter name:")
cv2.imwrite(filename="./images/" + n + ".jpg", img=opencvImage)
print("image saved!")
webcam.release()
cv2.waitKey(1650)
cv2.destroyAllWindows()
break
elif key == ord('q'):
print("off...")
webcam.release()
cv2.destroyAllWindows()
break
except (KeyboardInterrupt):
print("off...")
webcam.release()
cv2.destroyAllWindows()
break
def _prepare_detector(self):
"""Prepare the MTCNN detector.
This function should be invoked from a suitable Keras context
(with controlled TensorFlow Graph and Session), that is
usually it will be called via :py:meth:`run_tensorflow`.
"""
# Initialize the MTCNN detector
detector = mtcnn.MTCNN()
# The last part of the preparation process of MTCNN is to
# create the models' (P-net, R-net and O-net) predict
# functions. It may be possible to achieve this by calling
# model._make_predict_function() for each of them, but this is
# discourageds as it uses a private Keras API. The recomended
# way to create the predict function is to call predict, as
# the predict function will be automatically compiled on first
# invocation. Hence we provide some dummy image and invoke
# predict for all three networks by calling
# detector.detect_faces().
image = np.random.randint(0, 255, (200, 200, 3), np.uint8)
_ = detector.detect_faces(image)
self._detector = detector
def start_camera():
#cv2.namedWindow("preview")
vc = cv2.VideoCapture(0)
if vc.isOpened(): # try to get the first frame
rval, frame = vc.read()
else:
rval = False
while rval:
#cv2.imshow("preview", frame)
rval, frame = vc.read()
key = cv2.waitKey(20)
detector = mtcnn.MTCNN()
result = detector.detect_faces(frame)
if len(result) > 0:
image = crop_image(result, frame)
image_resize = cv2.resize(image, (224, 224))
image_reshape = np.reshape(image_resize, [1, 224, 224, 3])
prediction = model.predict(image_reshape)
#print(prediction)
if key == 27: # exit on ESC
break
def get_faces(file:str):
"""for file (path to file), open video and extract facial locations"""
j = 0
cap = cv2.VideoCapture(file)
length = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
faces = []
i = 0
detector = mtcnn.MTCNN()
while cap.isOpened():
ret, frame = cap.read()
i += 1
if ret==True:
print('file %d frame %1.2f' % (j, i/length))
f = detector.detect_faces(frame)
faces.append([x['box'] for x in f])
else:
break
print('writing to ' + file + 'faces.p')
pickle.dump(faces, open(file+'faces.p', 'wb'))
cap.release()
cv2.destroyAllWindows()
def extract_face(filename, required_size=None):
# Avoid the caveat of default collection argument
if required_size is None:
required_size = (160, 160)
image = Image.open(filename)
image = image.convert('RGB')
pixels = np.asarray(image)
detector = mtcnn.MTCNN()
results = detector.detect_faces(pixels)
x1, y1, width, height = results[0]['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
face = pixels[y1:y2, x1:x2]
image = Image.fromarray(face)
image = image.resize(required_size)
return np.asarray(image)
def __init__(self, face_cascade_path, eye_cascade_path,
shape_predictor_path):
self.cap = cv2.VideoCapture(0)
_, self.capture = self.cap.read()
self.face_cascade = cv2.CascadeClassifier(face_cascade_path)
self.eye_cascade = cv2.CascadeClassifier(eye_cascade_path)
self.pose_predictor = dlib.shape_predictor(shape_predictor_path)
self.mtcnn_detector = mtcnn.MTCNN()
self.hog_detector = dlib.get_frontal_face_detector()
self.current_state = {
"face": (0, 0, 100, 100),
"right_eye": (0, 0, 10, 10),
"left_eye": (0, 0, 10, 10),
"right_pupil": (0, 0),
"left_pupil": (0, 0),
"pose": [0 for i in range(0, 68 * 2)]
}
cv2.startWindowThread()
import cv2
import mtcnn
face_detector = mtcnn.MTCNN(min_face_size=50)
img = cv2.imread('Coffee.jpeg')
conf_t = 0.90
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
results = face_detector.detect_faces(img_rgb)
print(results)
for res in results:
x1, y1, width, height = res['box']
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
confidence = res['confidence']
if confidence < conf_t:
continue
key_points = res['keypoints'].values()
cv2.rectangle(img, (x1, y1), (x2, y2), (255, 255, 0), thickness=6)
# cv2.putText(img, f'conf: {confidence:.3f}', (x1, y1), cv2.FONT_ITALIC, 1, (0, 0, 255), 2)
for point in key_points:
cv2.circle(img, point, 4, (0, 255, 0), thickness=-1)
cv2.imshow('Sameer', img)
cv2.imwrite('Coffee_face_detected.jpeg', img)
cv2.waitKey(0)
def load_extractor():
global EXTRACTOR
if EXTRACTOR is None:
EXTRACTOR = mtcnn.MTCNN()
return EXTRACTOR
frame = res
if show:
cv2.imshow(win_name, frame)
if cv2.waitKey(1) & 0xff == ord(break_key):
break
# End
#-----------------------------------------------------------------------------
# build function of face detection and recognition
#-----------------------------------------------------------------------------
encoder_model = 'facenet_keras.h5'
required_size = (160, 160) # default size
face_detector = mtcnn.MTCNN() # set face detector
face_encoder = load_model(encoder_model)
# build a function to generate an embedding from single face
def get_embedding(photo):
'''
photo: path of photo file, only support jpg and png format
'''
# get features of face from photo
img = cv2.imread(photo)
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
results = face_detector.detect_faces(img_rgb)
# convert the features of face into embedding of model
embeddings = []
import cv2, mtcnn import numpy as np detect_net = mtcnn.MTCNN() imgpath = '../Data/Face_Normalization/' imglist = '../Data/Face_Normalization/crop_list.txt' savepath = '../Data/Face_Normalization/' boxes = detect_net.crop_image(imgpath, imglist, savepath)
import numpy as np
import cv2 as cv
import mtcnn
from matplotlib import pyplot as plt
detect_face = mtcnn.MTCNN()
cap = cv.VideoCapture(0)
while True:
ret, frame = cap.read()
frame_rgb = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
result = detect_face.detect_faces(frame_rgb)
print(frame_rgb.shape)
mask = np.zeros(frame_rgb.shape, np.uint8)
for res in result:
x1, y1, w, h = res['box']
x2, y2 = x1 + w, y1 + h
print("-------", res)
cv.rectangle(frame_rgb, (x1, y1), (x2, y2), (0, 0, 255), thickness=2)
mask[y1:y2, x1:x2, :] = 1
frame_rgb = frame * mask[:, :, 1, np.newaxis]
cv.imshow("Frame", frame_rgb)
def __init__(self):
self.model = mtcnn.MTCNN(min_face_size=20)
import tensorflow as tf
import mtcnn
export_path = './mtcnn'
model = mtcnn.MTCNN()
print(dir(model))
with tf.keras.backend.get_session() as sess:
tf.saved_model.simple_save(sess,
export_path,
inputs={'input_image': model.input},
outputs={t.name: t
for t in model.outputs})
def predict(filename, modelname):
app_root = os.path.dirname(os.path.abspath(__file__))
app_root = os.path.join(app_root, 'images')
print('# load the model')
model = load_model(os.path.join(app_root, 'facenet_keras.h5'))
print('loaded the model')
# load image from file
print('# load the file')
image = Image.open(os.path.join(app_root, filename))
print('# opened the flie')
# convert to RGB, if needed
image = image.convert('RGB')
# convert to array
pixels = np.asarray(image)
os.remove(os.path.join(app_root, filename))
# create the detector, using default weights
detector = mtcnn.MTCNN()
# detect faces in the image
results = detector.detect_faces(pixels)
# extract the bounding box from the first face
faces = list()
print('# load the model')
for i in range(0, len(results)):
x1, y1, width, height = results[i]['box']
# bug fix
x1, y1 = abs(x1), abs(y1)
x2, y2 = x1 + width, y1 + height
# extract the face
face = pixels[y1:y2, x1:x2]
# resize pixels to the model size
image = Image.fromarray(face)
image = image.resize((160, 160))
face_array = np.asarray(image)
faces.append(face_array)
#image.save(str(i)+".jpg")
#pyplot.subplot(2, 7, i+1)
#pyplot.axis('off')
#pyplot.imshow(face_array)
#pyplot.show()
facesarray = np.asarray(faces)
#getembeddings
newfaceX = list()
for face_pixels in facesarray:
embedding = get_embedding(model, face_pixels)
newfaceX.append(embedding)
newfaceX = np.asarray(newfaceX)
#print(newfaceX.shape)
#normalising
in_encoder = Normalizer(norm='l2')
newfaceX = in_encoder.transform(newfaceX)
loaded_model = pickle.load(
open(os.path.join(app_root, modelname + '.sav'), 'rb'))
aaa = loaded_model.predict(newfaceX)
data = load(os.path.join(app_root, modelname + 'encoder.npz'))
trainy, testy = data['arr_0'], data['arr_1']
out_encoder = LabelEncoder()
out_encoder.fit(trainy)
trainy = out_encoder.transform(trainy)
testy = out_encoder.transform(testy)
predict_names = out_encoder.inverse_transform(aaa)
return predict_names
#predict('file.jpg','ootypicsmodel.sav')
def __init__(self):
print("MTCNN version: " + mtcnn.__version__)
print("FaceDetection class initialized")
# create the detector, using default weights
self.detector = mtcnn.MTCNN()
import threading
import mtcnn
from flask import jsonify
import json
import requests
button_flag = [1,1,1,1,1,1]
feature_list = []
button_name = ['','evans','hermsworth','jeremy','mark','olsen']
input_img="abc.jpg"
modeldir = './model/20180402-114759.pb'
npy='./npy'
detector = mtcnn.MTCNN()
image_size = detector.image_size
input_image_size = detector.input_image_size
sess = tf.Session()
print('Loading Modal')
with sess.as_default():
facenet.load_model(modeldir)
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]
print('Start Recognition')
def __init__(self):
self.face_encoder, _ = get_network('FaceEncoder', 'test')
self.detector = mtcnn.MTCNN()
self.image_cnt = 0