def update(self):
"""
Update the real time prediction on the screen
"""
self.counter += 1
camera_open, frame = self.video.getFrame()
# process image here
if self.counter % 10 == 0:
self.counter = 0
image, predictions = detect_face(frame, self.model, self.detector,
self.predictor)
if len(predictions) > 0:
emotion = predictions[0]
# FER2013: 0 = Angry, 1 = Disgust, 2 = Fear, 3 = Happy, 4 = Sad, 5 = Surprise, 6 = Neutral
self.angerLabel["text"] = "angry:" + str(round(emotion[0], 3))
self.disgustLabel["text"] = "disgust:" + str(
round(emotion[1], 3))
self.fearLabel["text"] = "fear:" + str(round(emotion[2], 3))
self.happyLabel["text"] = "happy:" + str(round(emotion[3], 3))
self.sadnessLabel["text"] = "sad:" + str(round(emotion[4], 3))
self.surpriseLabel["text"] = "surprise:" + str(
round(emotion[5], 3))
self.neutralLabel["text"] = "neutral:" + str(
round(emotion[6], 3))
if camera_open:
self.cur_frame = ImageTk.PhotoImage(image=Image.fromarray(frame))
self.canvas.create_image(0, 0, image=self.cur_frame, anchor=NW)
self.window.after(30, self.update) # rate to call update function
def snapshot(self):
"""
used to capture the image with emoji
"""
captured, frame = self.video.getFrame()
frame, predictions = detect_face(frame, self.model, self.detector,
self.predictor)
if captured:
image = "IMG-" + time.strftime("%H-%M-%S-%d-%m") + ".jpg"
cv.imwrite(image, cv.cvtColor(frame, cv.COLOR_BGR2RGB))
Label(self.window, text="image saved").place(x=430, y=510)
def get_stats(self):
"""
Get the predicted emotion
"""
camera_open, frame = self.video.getFrame()
image, predictions = detect_face(frame, self.model, self.detector,
self.predictor)
if len(predictions) > 0:
emotion = predictions[0]
# FER2013: 0 = Angry, 1 = Disgust, 2 = Fear, 3 = Happy, 4 = Sad, 5 = Surprise, 6 = Neutral
self.angerLabel["text"] = "angry:" + str(emotion[0])
self.disgustLabel["text"] = "disgust:" + str(emotion[1])
self.fearLabel["text"] = "fear:" + str(emotion[2])
self.happyLabel["text"] = "happy:" + str(emotion[3])
self.sadnessLabel["text"] = "sad:" + str(emotion[4])
self.surpriseLabel["text"] = "surprise:" + str(emotion[5])
self.neutralLabel["text"] = "neutral:" + str(emotion[6])
print('Read data dimension: ', img.ndim)
except (IOError, ValueError, IndexError) as e:
errorMessage = '{}: {}'.format(image_path, e)
print(errorMessage)
else:
if img.ndim < 2:
print('Error! Unable to align "%s"' % image_path)
text_file.write('%s\n' % (output_filename))
continue
if img.ndim == 2:
img = facenet.to_rgb(img)
print('to_rgb data dimension: ', img.ndim)
img = img[:, :, 0:3]
print('After data dimension: ', img.ndim)
bounding_boxes, _ = detect_face.detect_face(
img, minsize, pnet, rnet, onet, threshold, factor)
nrof_faces = bounding_boxes.shape[0]
print('Number of Detected Face(s): %d' % nrof_faces)
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)
def detect(frame, pnet, rnet, onet, faceTrackers):
big_image = frame.copy()
big_height = big_image.shape[0]
big_width = big_image.shape[1]
small_image = imutils.resize(frame, width=450)
small_image = cv2.cvtColor(small_image, cv2.COLOR_BGR2GRAY)
small_image = np.dstack([small_image, small_image, small_image])
small_height = small_image.shape[0]
small_width = small_image.shape[1]
gray = cv2.cvtColor(small_image, cv2.COLOR_BGR2GRAY)
minsize = 20 # minimum size of face
threshold = [0.6, 0.7, 0.7] # three steps's threshold
factor = 0.709
if gray.ndim == 2:
img = to_rgb(gray)
bounding_boxes, _ = detect_face.detect_face(img, minsize, pnet, rnet, onet,
threshold, factor)
nrof_faces = bounding_boxes.shape[0]
for face_position in bounding_boxes:
face_position = face_position.astype(int)
left_top = get_point(face_position[0], face_position[1], small_width,
small_height, big_width, big_height)
right_bottom = get_point(face_position[2], face_position[3],
small_width, small_height, big_width,
big_height)
x = left_top[0]
y = left_top[1]
w = right_bottom[0] - left_top[0]
h = right_bottom[1] - left_top[1]
x_bar = x + 0.5 * w
y_bar = y + 0.5 * h
matchedFid = None
for fid in faceTrackers.keys():
tracked_position = faceTrackers[fid].get_position()
t_x = int(tracked_position.left())
t_y = int(tracked_position.top())
t_w = int(tracked_position.width())
t_h = int(tracked_position.height())
t_x_bar = t_x + 0.5 * t_w
t_y_bar = t_y + 0.5 * t_h
if ((t_x <= x_bar <= (t_x + t_w)) and (t_y <= y_bar <= (t_y + t_h))
and (x <= t_x_bar <= (x + w)) and (y <= t_y_bar <=
(y + h))):
matchedFid = fid
if matchedFid is None:
print("Creating new tracker ")
tracker = dlib.correlation_tracker()
tracker.start_track(
big_image,
dlib.rectangle(x - 20, y - 20, x + w + 20, y + h + 20))
faceTrackers[len(faceTrackers)] = tracker
return (faceTrackers, big_image)
def identify(self, frame):
# Force frame to have RGB channels
if frame.ndim == 2:
frame = facenet.to_rgb(frame)
frame = frame[:, :, 0:3]
# Get the bounding boxes
bounding_boxes, _ = detect_face.detect_face(frame, self.minsize,
self.pnet, self.rnet,
self.onet, self.threshold,
self.factor)
# Get the number of faces detected
nrof_faces = bounding_boxes.shape[0]
if nrof_faces > 1:
print('[AUTH_ERROR] Multiple equidistant faces detected!')
print('[AUTH_ERROR] Please maintain one-by-one queue!')
return 1
elif nrof_faces == 1:
bb = np.zeros((nrof_faces, 4), dtype=np.int32)
bb[0][0] = bounding_boxes[0][0]
bb[0][1] = bounding_boxes[0][1]
bb[0][2] = bounding_boxes[0][2]
bb[0][3] = bounding_boxes[0][3]
# Get the frame size
img_size = np.asarray(frame.shape)[0:2]
# For storing cropped, scaled and scaled+reshaped image
cropped = None
scaled = None
scaled_reshape = None
# Create Embedding array
emb_array = np.zeros((1, self.embedding_size))
# Bounding box out of frame size range exception
if bb[0][0] <= 0 or bb[0][1] <= 0 or bb[0][2] >= len(
frame[0]) or bb[0][3] >= len(frame[1]):
print('[ERROR] Bounding Box out of frame size range!')
return 2
try:
cropped = frame[bb[0][1]:bb[0][3], bb[0][0]:bb[0][2], :]
cropped = facenet.flip(cropped, False)
scaled = np.array(
Image.fromarray(cropped).resize(
(self.image_size, self.image_size),
resample=Image.BILINEAR))
scaled = cv2.resize(
scaled, (self.input_image_size, self.input_image_size),
interpolation=cv2.INTER_CUBIC)
scaled = facenet.prewhiten(scaled)
scaled_reshape = scaled.reshape(-1, self.input_image_size,
self.input_image_size, 3)
feed_dict = {
self.images_placeholder: scaled_reshape,
self.phase_train_placeholder: False
}
emb_array[0, :] = self.sess.run(self.embeddings,
feed_dict=feed_dict)
predictions = self.model.predict_proba(emb_array)
# print('predictions:', predictions)
best_class_indices = np.argmax(predictions, axis=1)
# print('best_class_indices:', best_class_indices)
best_class_probabilities = predictions[
np.arange(len(best_class_indices)), best_class_indices]
# print('best_class_probabilities:', best_class_probabilities)
except Exception as e:
print('[ERROR]', e)
return 3
if best_class_probabilities[0] > 0.7:
return (self.faceIds[best_class_indices[0]], bb,
best_class_probabilities[0])
else:
print('[AUTH_ERROR] ACCESS DENIED!')
return 4
else:
print('[INFO] No detected face in the threshold vicinity!')
return 5
def mtcnn_simplify(image_path, image_size, margin, gpu_memory_fraction,
detect_multiple_faces):
sleep(random.random())
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 = detect_face.create_mtcnn(sess, None)
minsize = 20 # minimum size of face
threshold = [0.6, 0.7, 0.7] # three steps's threshold
factor = 0.709 # scale factor
try:
img = misc.imread(image_path)
except (IOError, ValueError, IndexError) as e:
errorMessage = '{}: {}'.format(image_path, e)
print(errorMessage)
else:
if img.ndim < 2:
print('Unable to align "%s"' % image_path)
os._exit()
if img.ndim == 2:
img = function.to_rgb(img)
img = img[:, :, 0:3]
bounding_boxes, _ = detect_face.detect_face(img, minsize, pnet, rnet,
onet, threshold, factor)
nrof_faces = bounding_boxes.shape[0]
# --------需要修改，建立一个文件夹用来储存无法剪裁的文件
with open('E:/Desktop/unableface.txt', 'w') as text:
if nrof_faces > 0:
det = bounding_boxes[:, 0:4]
det_arr = []
img_size = np.asarray(img.shape)[0:2]
det_arr.append(np.squeeze(det))
for i, det in enumerate(det_arr):
det = np.squeeze(det)
bb = np.zeros(4, dtype=np.int32)
bb[0] = np.maximum(det[0] - margin / 2, 0)
bb[1] = np.maximum(det[1] - margin / 2, 0)
bb[2] = np.minimum(det[2] + margin / 2, img_size[1])
bb[3] = np.minimum(det[3] + margin / 2, img_size[0])
cropped = img[bb[1]:bb[3], bb[0]:bb[2], :]
# scaled 就是检测对齐剪裁过后的结果
img_mtcnn = misc.imresize(cropped,
(image_size, image_size),
interp='bilinear')
# ---------需要修改，为储存的目录
image_new_name = 'E:' + image_path.split(':')[-1].split(
'.')[0] + '.jpg'
image_new_path = os.path.dirname(image_new_name)
mkdir(image_new_path)
cv2.imwrite(image_new_name, img_mtcnn)
# print(image_new_name)
else:
print('unable to align"%s"' % image_path)
text.write('%s\n' % (image_path))
def monitor(refreshTime):
global count
#1. Recognise the person using the computer, implement for multiple people later
user_id, seconds = detect_face()
if user_id == None: # No one in front of the computer --> lock or standby
count = count+1
if count > 3:
os.popen('gnome-screensaver-command --lock')
count = 0
displayText = "No user found"
elif user_id == -1: # An anonymous user is sitting in front of the computer --> Prevent keyboard, mouse input
displayText = "Anonymous user"
else: # The face detector has detected a person properly --> Regain mouse keyboard input, increment screen time for applications
count = 0
# get open apps
open_apps = get_win_info()
displayText, user, apps, isUnderAge, responseStatus = currentUserDetails(user_id)
# verify open apps
if (responseStatus == 200):
print("Detected " + user)
for app in apps:
for open_app in open_apps:
pid, process_name, title = open_app
if (app['title'].lower() == process_name):
print("Observing " + title)
print("Time spent on " + app['title'] + " is " + str(app['time_today']) + ' seconds')
print("****************************************************************************")
increment(app, seconds + refreshTime)
profanity = check_profanity(title)
if (profanity != False and isUnderAge == True): # If age restricted applications used by underage children at home
notif('AGE RESTRICTED CONTENT!!!', f'{profanity}. Your action will be reported!')
set_warning(app, profanity)
toggle_process(pid, 2)
break
# time limit warning
remaining_time = app['time_limit'] - app['time_today']
print(remaining_time)
if (remaining_time > 2*refreshTime and remaining_time < 3*refreshTime ):
message = 'The time limit of ' + app['title'] + ' will be reached after ' + str(remaining_time) + ' seconds.'
notif('Session limit warning', message)
# time limit ended
if (remaining_time <= 0):
message = 'The time limit of ' + app['title'] + ' has been reached.'
notif('Time limit ended', message)
toggle_process(pid, 1)
else:
toggle_process(pid, 0)
return displayText