def run(self):
y_labels = []
x_train = []
cur_id = 0
label_ids = {}
# fetching images from dataset for training
for root, dirs, files in os.walk(self.face_images_dataset_dir):
# FIXME - adding talkative settings in prefs !!!
# if our robot it too talkative, emit this signal
g_emitter().emit_signal_to_feed_mouth("checking %s" %
os.path.basename(root))
for file in files:
# check file extension for image files
extension = os.path.splitext(file)[1]
if extension in [".jpg", ".jpeg", ".png"]:
full_path = os.path.join(root, file)
label = os.path.basename(root).replace(" ", "-").lower()
if label not in label_ids:
label_ids[label] = cur_id
cur_id += 1
img_id = label_ids[label]
log.debug("FaceTrainer :: %s - %s - %s" %
(str(label), str(img_id), str(full_path)))
self.processing_image.emit(label, full_path)
# convert image to grayscale
pil_image = Image.open(full_path).convert("L")
# convery grayscale image to numpy array
image_array = np.array(pil_image, "uint8")
faces = self.face_cascade.detectMultiScale(
image_array, 1.3, 5)
for (x, y, w, h) in faces:
# define roi for eyes detection,ideally,
# we should detect eyes within the rectangular
# bounds of a face
roi = image_array[y:y + h, x:x + w]
x_train.append(roi)
y_labels.append(img_id)
# save trained labels
with open("dataset/face_trainer_labels.pickle", 'wb') as f:
pickle.dump(label_ids, f)
self.recognizer.train(x_train, np.array(y_labels))
self.recognizer.save("dataset/face_trainer.yml")
self.face_training_finished.emit()
def run(self):
while True:
if not self.queue.empty():
text = self.queue.get()
log.info("Mouth speaking text: %s" % text)
# ignore empty/None texts
if not text or not len(text):
continue
speak_text(text)
# tell face to change mouth animations to speaking
g_emitter().emit_signal_to_set_speaking_state()
time.sleep(.1)
else:
# tell face to change mouth animations to idle
time.sleep(.2)
g_emitter().emit_signal_to_set_idle_state()
def __init__(self):
QWidget.__init__(self)
# loaind ui from xml
uic.loadUi(os.path.join(DIRPATH, 'app.ui'), self)
# FIXME - libpng warning: iCCP: known incorrect sRGB profile
self.setWindowIcon(QIcon("./images/robot_icon.png"))
# keep the window fixed sized
self.setFixedSize(self.size())
# button event handlers
self.btnStartCaptureForVideoAnalysis.clicked.connect(
self.start_capture_for_video_analysis)
self.btnStopCaptureForVideoAnalysis.clicked.connect(
self.stop_capture_for_video_analysis)
self.btnChooseClassifierXML.clicked.connect(
self.choose_classifier_file)
self.btnChooseImage.clicked.connect(self.choose_image_for_analysis)
self.setup_tray_menu()
# add camera ids
for i in range(0, 11):
self.cboxCameraIds.addItem(str(i))
self.cboxCameraIds1.addItem(str(i))
# setting up handlers for menubar actions
self.actionAbout.triggered.connect(self.about)
self.actionExit.triggered.connect(qApp.quit)
self.actionPreferences.triggered.connect(self.show_preferences)
# video analysis image widget
self.img_widget_vid_analysis = ImageWidget()
self.hlayoutVideoAnalysis.addWidget(self.img_widget_vid_analysis)
# face training image widget
self.img_widget_face_training = ImageWidget()
self.hlayoutFaceTrainingImg.addWidget(self.img_widget_face_training)
# face identification image widget
self.img_widget_identify_face = ImageWidget()
self.hlayoutIdentifyFace.addWidget(self.img_widget_identify_face)
# image analysis image widget
self.img_widget_img_analysis = ImageWidget()
self.hlayoutImageAnalysis.addWidget(self.img_widget_img_analysis)
img = cv2.imread("images/human.png")
self.img_widget_img_analysis.handle_image_data(img)
self.vid_capture = VideoCapture()
self.vid_capture.got_image_data_from_camera.connect(
self.process_image_data_from_camera)
self.highlight_faces = self.chkHighlightFaces.isChecked()
self.chkHighlightFaces.stateChanged.connect(
self.highlight_faces_checkbox_changed)
self.chckGrayscale.stateChanged.connect(
self.grayscale_checkbox_changed)
# face trainer dataset browser btn handler
self.btnBrowseDatasetForFaceTrainer.clicked.connect(
self.browse_dataset_for_face_trainer)
self.btnBrowseClassifierForFaceTrainer.clicked.connect(
self.browse_classifier_file_for_face_trainer)
self.btnStartFaceTrainer.clicked.connect(self.start_face_trainer)
self.btnBrowseIdentifyFace.clicked.connect(self.browse_identify_face)
self.btnTalk.clicked.connect(self.lets_talk)
# create and start robot
self.robot = Robot(self.lblRobot)
self.mouth = Mouth()
# connect global signals to slots
g_emitter().feed_mouth.connect(self.mouth.feed_text)
g_emitter().set_speaking_state.connect(self.robot.set_speaking_state)
g_emitter().set_idle_state.connect(self.robot.set_idle_state)
self.robot.start()
self.mouth.start()
def face_training_finished(self):
self.lblFaceTrainingStatus.setText("FACE TRAINING FINISHED")
g_emitter().emit_signal_to_feed_mouth("face training finished")
def lets_talk(self):
text = self.teTalk.toPlainText()
self.teTalk.setText("")
g_emitter().emit_signal_to_feed_mouth(text)
