def handleNewRoi(self, roi, width, height):
roi = cv.resize(roi, (96, 96))
images = np.array([roi])
images = AgenderNetMobileNetV2.prep_image(images)
prediction = self.model.predict(images)
genders, ages = AgenderNetMobileNetV2.decode_prediction(prediction)
print(genders, ages)
if len(genders) > 0:
return genders[0], ages[0]
return None, None
def __init__(self, db, *args, **kwargs):
"""
The main window of the program
"""
# initialise db
self.db = db
# sets the frame and the pages
tk.Frame.__init__(self, *args, **kwargs)
self.p1 = TablePage(db, self)
self.p2 = CamPage(self)
self.p3 = DisplayPage(self)
# sets a button frame at the top, insert into the window
buttonframe = tk.Frame(self)
container = tk.Frame(self)
buttonframe.pack(side="top", fill="x", expand=False)
container.pack(side="top", fill="both", expand=True)
# place the button at the top
self.p1.place(in_=container, x=0, y=0, relwidth=1, relheight=1)
self.p2.place(in_=container, x=0, y=0, relwidth=1, relheight=1)
self.p3.place(in_=container, x=0, y=0, relwidth=1, relheight=1)
# set the function of each button
b1 = tk.Button(buttonframe, text="Adverts", command=self.p1.lift)
b2 = tk.Button(buttonframe, text="Cam", command=self.p2.lift)
b3 = tk.Button(buttonframe, text="Display", command=self.p3.lift)
# position the buttons
b1.pack(side="left")
b2.pack(side="left")
b3.pack(side="left")
# show the first page first
self.p1.show()
# values for to be used in the few functions
width, height = 600, 600
# captured video object with height and width wanted
self.cap = cv2.VideoCapture(0)
self.cap.set(cv2.CAP_PROP_FRAME_WIDTH, width)
self.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, height)
# sets margin and image size
self.margin = 0.4
self.img_size = 96
# gets the face detector
self.detector = dlib.get_frontal_face_detector()
# loads the model and weights
self.model = AgenderNetMobileNetV2()
self.model.load_weights(
'model/weight/mobilenetv2/mobilenet_v2-02-4.0368-0.8085-9.6728.h5')
# calls the function to show the frame
self.show_frame()
def main():
args = parser.parse_args()
MODEL = args.model
model = None
logger.info('Load model and weight')
if MODEL == 'mobilenetv2':
model = AgenderNetMobileNetV2()
model.load_weights(
'model/weight/mobilenetv2/model.10-3.8290-0.8965-6.9498.h5')
elif MODEL == 'inceptionv3':
model = AgenderNetInceptionV3()
model.load_weights(
'model/weight/inceptionv3/model.16-3.7887-0.9004-6.6744.h5')
elif MODEL == 'mobilenetv1':
model = AgenderNetMobileNetV1()
#model.load_weigts('model/weight/mobilenetv1/model.h5')
else:
model = AgenderSSRNet(64, [3, 3, 3], 1.0, 1.0)
model.load_weights(
'model/weight/ssrnet/model.37-7.3318-0.8643-7.1952.h5')
logger.info('Read image')
image = cv2.imread(
'data/imdb_aligned/02/nm0000002_rm1346607872_1924-9-16_2004.jp')
image = cv2.resize(image, (model.input_size, model.input_size))
image = np.expand_dims(image, axis=0)
image = model.prep_image(image)
logger.info('Predict with {}'.format(MODEL))
wrapped = wrapper(predictone, model, image)
logger.info(proces_time(wrapped))
def main():
args = parser.parse_args()
DB = args.db_name
MODEL = args.model
data = pd.read_csv('data/db/{}.csv'.format(DB))
model = None
logger.info('Load model and weight')
if MODEL == 'mobilenetv2':
model = AgenderNetMobileNetV2()
model.load_weights(
'model/weight/mobilenetv2/model.10-3.8290-0.8965-6.9498.h5')
elif MODEL == 'inceptionv3':
model = AgenderNetInceptionV3()
model.load_weights(
'model/weight/inceptionv3/model.16-3.7887-0.9004-6.6744.h5')
elif MODEL == 'mobilenetv1':
model = AgenderNetMobileNetV1()
#model.load_weights('model/weight/mobilenetv1/___.h5')
else:
model = AgenderSSRNet(64, [3, 3, 3], 1.0, 1.0)
model.load_weights(
'model/weight/ssrnet/model.37-7.3318-0.8643-7.1952.h5')
logger.info('Read image')
images = [
cv2.imread('{}_aligned/{}'.format(DB, path))
for path in tqdm(data.full_path.values)
]
images = [
cv2.resize(image, (model.input_size, model.input_size))
for image in images
]
images = np.array(images)
images = model.prep_image(images)
logger.info('Predict data')
start = time.time()
prediction = model.predict(images)
pred_gender, pred_age = model.decode_prediction(prediction)
elapsed = time.time() - start
logger.info('Time elapsed {:.2f} sec'.format(elapsed))
result = pd.DataFrame()
result['full_path'] = data['full_path']
result['age'] = pred_age
result['gender'] = pred_gender
result.to_csv('result/{}.csv'.format(DB), index=False)
def __init__(self):
self.model = AgenderNetMobileNetV2()
self.model.load_weights(
'model/weight/mobilenetv2/model.10-3.8290-0.8965-6.9498.h5')
print("DemographicDetector() -> __init__ complete!")
def main():
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
sess = tf.Session(config=config)
K.tensorflow_backend.set_session(sess)
args = parser.parse_args()
MODEL = args.model
EPOCH = args.epoch
BATCH_SIZE = args.batch_size
NUM_WORKER = args.num_worker
db, paths, age_label, gender_label = load_data()
n_fold = 1
print('[K-FOLD] Started...')
kf = KFold(n_splits=10, shuffle=True, random_state=1)
kf_split = kf.split(age_label)
for train_idx, test_idx in kf_split:
model = None
if MODEL == 'ssrnet':
model = AgenderSSRNet(64, [3, 3, 3], 1.0, 1.0)
elif MODEL == 'inceptionv3':
model = AgenderNetInceptionV3()
elif MODEL == 'mobilenetv1':
model = AgenderNetMobileNetV1()
else:
model = AgenderNetMobileNetV2()
train_db = db[train_idx]
train_paths = paths[train_idx]
train_age = age_label[train_idx]
train_gender = gender_label[train_idx]
test_db = db[test_idx]
test_paths = paths[test_idx]
test_age = age_label[test_idx]
test_gender = gender_label[test_idx]
losses = {
"age_prediction": "categorical_crossentropy",
"gender_prediction": "categorical_crossentropy",
}
metrics = {
"age_prediction": mae,
"gender_prediction": "acc",
}
if MODEL == 'ssrnet':
losses = {
"age_prediction": "mae",
"gender_prediction": "mae",
}
metrics = {
"age_prediction": "mae",
"gender_prediction": "binary_accuracy",
}
callbacks = [
ModelCheckpoint(
"train_weight/{}-{epoch:02d}-{val_loss:.4f}-{val_gender_prediction_acc:.4f}-{val_age_prediction_mae:.4f}.h5"
.format(MODEL),
verbose=1,
save_best_only=True,
save_weights_only=True),
CSVLogger('train_log/{}-{}.log'.format(MODEL, n_fold))
]
if MODEL == 'ssrnet':
callbacks = [
ModelCheckpoint(
"train_weight/{}-{epoch:02d}-{val_loss:.4f}-{val_gender_prediction_binary_accuracy:.4f}-{val_age_prediction_mean_absolute_error:.4f}.h5"
.format(MODEL),
verbose=1,
save_best_only=True,
save_weights_only=True),
CSVLogger('train_log/{}-{}.log'.format(MODEL, n_fold)),
DecayLearningRate([30, 60])
]
model.compile(optimizer='adam', loss=losses, metrics=metrics)
model.fit_generator(DataGenerator(model, train_db, train_paths,
train_age, train_gender, BATCH_SIZE),
validation_data=DataGenerator(
model, test_db, test_paths, test_age,
test_gender, BATCH_SIZE),
epochs=EPOCH,
verbose=2,
workers=NUM_WORKER,
use_multiprocessing=True,
max_queue_size=int(BATCH_SIZE * 2),
callbacks=callbacks)
n_fold += 1
del train_db, train_paths, train_age, train_gender
del test_db, test_paths, test_age, test_gender
mot_tracker.update(dets, genders, ages)
for tracker in mot_tracker.trackers:
(left, top, right, bottom) = convert_x_to_bbox(
tracker.kf.x[:4, :]).astype('int').flatten()
cv2.rectangle(frame, (left, top), (right, bottom), (0, 255, 0), 2)
age = tracker.smooth_age()
gender = 'M' if tracker.smooth_gender() == 1 else 'F'
cv2.putText(frame, "id: {} {} {}".format(tracker.id, gender, age),
(left - 10, top - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.45, (0, 255, 0), 2)
cv2.putText(frame, "{:.1f} FPS".format(fps.fps()), (1100, 50),
cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 0), 2)
cv2.namedWindow("Frame", cv2.WINDOW_NORMAL)
cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF
fps.update()
grabbed, frame = stream.read()
fps.stop()
stream.release()
cv2.destroyAllWindows()
if __name__ == '__main__':
print('[INFO] Load model')
model = AgenderNetMobileNetV2()
print('[INFO] Load weight')
model.load_weights(
'model/weight/mobilenetv2/model.10-3.8290-0.8965-6.9498.h5')
graph = tf.get_default_graph()
main()