def process_video():
"""Process video and handle user commands."""
# Feed from computer camera with threading
cap = video.VideoStream(src=0, resolution=(200, 150)).start()
# Define the codec and create VideoWriter object
fourcc = cv2.VideoWriter_fourcc(*'mp4v') # Be sure to use the lower case
out = cv2.VideoWriter('output.mp4', fourcc, 20.0, (200, 150))
while True:
# Read image, anonymize, and wait for user quit.
image = cap.read()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
nb_rects = anonymizer(image, gray)
image = cv2.convertScaleAbs(image, alpha=1.0, beta=0)
if nb_rects > 3:
image = cv2.putText(image, 'GOOD', (10, 500), font, 5,
(255, 255, 255), 4, cv2.LINE_AA)
else:
image = cv2.putText(image, 'BAD', (10, 500), font, 5, (0, 0, 255),
4, cv2.LINE_AA)
out.write(image)
cv2.imshow("Realtime Anon", image)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
out.release()
cv2.destroyAllWindows()
cap.stop()
def facial_landmarks():
# Feed from computer camera with threading
cap = video.VideoStream(src=0).start()
while True:
# Getting out image by webcam
image = cap.read()
# Converting the image to gray scale
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
if face_detection_method == 'hog':
hog_detector(image, gray)
else:
dl_detector(image, gray)
# show the output frame
# adding brightness and contrast -> α⋅p(i,j)+β where p(i.j) is pixel value for each point
image = cv2.convertScaleAbs(image, alpha=1.0, beta=0)
cv2.imshow("Facial Landmarks", image)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cv2.destroyAllWindows()
cap.stop()
def stream(camera_url):
print('[INFO] starting video stream...')
if camera_url == None:
print('[ERROR] video .....')
vs = video.VideoStream(camera_url).start()
vd = stream_processing(vs)
return vd
def face_detection_realtime():
# Feed from computer camera with threading
# total = video.count_frames(video_patpyth)
cap = video.VideoStream(0).start()
# cap = cv2.VideoCapture(video_path)
count = 0
while True:
try:
count += 1
print(f"{count}")
# Getting out image frame by webcam
# _, img = cap.read()
img = cap.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# https://docs.opencv.org/trunk/d6/d0f/group__dnn.html#ga29f34df9376379a603acd8df581ac8d7
inputBlob = cv2.dnn.blobFromImage(cv2.resize(img, (300, 300)), 1,
(300, 300), (104, 177, 123))
detector.setInput(inputBlob)
detections = detector.forward()
img, age, gender = find_faces(img, detections, gray)
# return stream video to client
imgencode = cv2.imencode('.jpg', img)[1]
stringData = imgencode.tostring()
yield (b'--frame\r\n'
b'Content-Type: text/plain\r\n\r\n' + stringData + b'\r\n')
yield age
yield gender
# cv2.imshow('img', img)
# if cv2.waitKey(1) & 0xFF == ord("q"):
# break
except:
pass
del (cap)
cv2.destroyAllWindows()
cap.stop()
def face_detection_realtime():
# Feed from computer camera with threading
cap = video.VideoStream(src=0).start()
while True:
# Getting out image frame by webcam
img = cap.read()
# https://docs.opencv.org/trunk/d6/d0f/group__dnn.html#ga29f34df9376379a603acd8df581ac8d7
inputBlob = cv2.dnn.blobFromImage(cv2.resize(img, (300, 300)), 1,
(300, 300), (104, 177, 123))
detector.setInput(inputBlob)
detections = detector.forward()
find_faces(img, detections)
if cv2.waitKey(1) & 0xFF == ord("q"):
break
cv2.destroyAllWindows()
cap.stop()
def process_video():
"""Process video and handle user commands."""
# Feed from computer camera with threading
cap = video.VideoStream(src=0, resolution=(200, 150)).start()
while True:
# Read image, anonymize, and wait for user quit.
image = cap.read()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
nb_rects = anonymizer(image, gray)
image = cv2.convertScaleAbs(image, alpha=1.0, beta=0)
if nb_rects > 5:
image = cv2.putText(image, 'GOOD', (10, 500), font, 5,
(255, 255, 255), 4, cv2.LINE_AA)
else:
image = cv2.putText(image, 'BAD', (10, 500), font, 5, (0, 0, 255),
4, cv2.LINE_AA)
cv2.imshow("Realtime Anon", image)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cv2.destroyAllWindows()
cap.stop()
def run(self):
print("started facial recognition")
vs = video.VideoStream(src=0).start()
match_found = False
while not match_found and self.heartbeat():
frame = vs.read()
# raspberry pi camera sensor is flipped 180
frame = cv2.rotate(frame, cv2.ROTATE_180)
# convert from bgr to rgb then shrink to speed up detection
rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
rgb = imutils.resize(rgb, width=self.RESIZE)
# get bounding box coordinates for each face in image
boxes = face_recognition.face_locations(
rgb, model=self.DETECTION_METHOD)
# compute facial embeddings for each face
encodings = face_recognition.face_encodings(rgb, boxes)
# look for user in faces found in video frame
for encoding in encodings:
# find how many user face encodings match with the captured face
matches = face_recognition.compare_faces(
self.user_encodings, encoding)
perc = matches.count(True) / len(matches)
print("{} % match for user found".format(perc * 100))
# matches with atleast 25% of user encodings
if perc >= 0.25:
# match found
print("sending unlock request")
match_found = True
break
if match_found:
# execute callback
self.callback()
self.killed()
# that are viewing the stream)
output_frame = None
lock = threading.Lock()
# Get the relative path to this file (we will use it later)
FILE_PATH = os.path.dirname(os.path.realpath(__file__))
# initialize the facial recognizer
f_r = facial_recognition.FacialRecognizer()
# * ---------- Create App --------- *
app = Flask(__name__)
CORS(app, support_credentials=True)
# Initialize the video stream
video_stream = video.VideoStream(src=0).start()
time.sleep(2.0)
# # * ---------- DATABASE CONFIG --------- *
# DATABASE_USER = os.environ['DATABASE_USER']
# DATABASE_PASSWORD = os.environ['DATABASE_PASSWORD']
# DATABASE_HOST = os.environ['DATABASE_HOST']
# DATABASE_PORT = os.environ['DATABASE_PORT']
# DATABASE_NAME = os.environ['DATABASE_NAME']
def DATABASE_CONNECTION():
try:
return psycopg2.connect(
user="******",
password="",