def __init__(self, LogPath):
""" Initializes the Facenet class. """
# Class settings
self.Known = []
self.Helpers = Helpers()
self.LogFile = self.Helpers.setLogFile(self.Helpers.confs["Server"]["Logs"]+"/Facenet")
# OpenCV settings
self.OpenCV = OpenCV(self.Helpers)
# Dlib settings
self.Detector = dlib.get_frontal_face_detector()
self.Predictor = dlib.shape_predictor(self.Helpers.confs["Classifier"]["Dlib"])
def connect(self):
""" Connects to the Foscam IP camera. """
self.camera = OpenCV("rtsp://" +
self.Helpers.confs["Foscam"]["RTSPuser"] + ":" +
self.Helpers.confs["Foscam"]["RTSPpass"] + "@" +
self.Helpers.confs["Foscam"]["RTSPip"] + ":" +
self.Helpers.confs["Foscam"]["RTSPport"] + "/" +
self.Helpers.confs["Foscam"]["RTSPendpoint"])
self.Helpers.logger.info("Connected To Camera")
def __init__(self, LogPath, Configuration):
###############################################################
#
# Sets up all default requirements and placeholders
#
###############################################################
# Class settings
self.Known = []
self.Configuration = Configuration
self.Helpers = Helpers()
self.LogFile = self.Helpers.setLogFile(
self.Configuration.AiCore["Logs"] + "/Facenet")
# OpenCV settings
self.OpenCV = OpenCV(self.Helpers)
# Dlib settings
self.Detector = dlib.get_frontal_face_detector()
self.Predictor = dlib.shape_predictor(
self.Configuration.Classifier["Dlib"])
def __init__(self):
""" Initializes the AML/ALL Detection System Movidius NCS1 Classifier Class. """
self.Helpers = Helpers("ClassifierServer")
self.confs = self.Helpers.confs
self.OpenCV = OpenCV()
self.Movidius = Movidius()
self.Movidius.checkNCS()
self.Movidius.loadInception()
self.Helpers.logger.info(
"Classifier server class initialization complete.")
class Facenet():
###############################################################
#
# Facenet helper functions
#
###############################################################
def __init__(self, LogPath, Configuration):
###############################################################
#
# Sets up all default requirements and placeholders
#
###############################################################
# Class settings
self.Known = []
self.Configuration = Configuration
self.Helpers = Helpers()
self.LogFile = self.Helpers.setLogFile(
self.Configuration.AiCore["Logs"] + "/Facenet")
# OpenCV settings
self.OpenCV = OpenCV(self.Helpers)
# Dlib settings
self.Detector = dlib.get_frontal_face_detector()
self.Predictor = dlib.shape_predictor(
self.Configuration.Classifier["Dlib"])
def PreprocessKnown(self, ValidDir, Graph):
###############################################################
#
# Preprocesses known images
#
###############################################################
for validFile in os.listdir(ValidDir):
if os.path.splitext(validFile)[1] in self.Configuration.Classifier[
"ValidIType"]:
self.Known.append({
"File":
validFile,
"Score":
self.Infer(
cv2.resize(cv2.imread(ValidDir + validFile),
(640, 480)), Graph)
})
self.Helpers.logMessage(self.LogFile, "Facenet", "STATUS",
str(len(self.Known)) + " known images found.")
def ProcessFrame(self, Frame):
###############################################################
#
# Processes frame
#
###############################################################
Known = []
Frame = cv2.resize(cv2.imdecode(Frame, cv2.IMREAD_UNCHANGED),
(640, 480))
RawFrame = Frame.copy()
Gray = cv2.cvtColor(Frame, cv2.COLOR_BGR2GRAY)
Path = "Data/Captured/" + datetime.now().strftime(
"%Y-%m-%d") + "/" + datetime.now().strftime("%H") + "/"
FileName = datetime.now().strftime('%M-%S') + ".jpg"
FileNameGray = datetime.now().strftime('%M-%S') + "-Gray.jpg"
self.OpenCV.SaveFrame(Path + "/", FileName, Frame)
self.OpenCV.SaveFrame(Path + "/", FileNameGray, Gray)
return Frame
def LoadGraph(self):
###############################################################
#
# Loads Facenet graph
#
###############################################################
with open(self.Configuration.Classifier["NetworkPath"] +
self.Configuration.Classifier["Graph"],
mode='rb') as f:
graphFile = f.read()
self.Helpers.logMessage(self.LogFile, "Facenet", "Status",
"Loaded TASS Graph")
return graphFile
def Infer(self, face, graph):
###############################################################
#
# Runs the image through Facenet
#
###############################################################
graph.LoadTensor(self.PreProcess(face).astype(np.float16), None)
output, userobj = graph.GetResult()
return output
def PreProcess(self, src):
###############################################################
#
# Preprocesses an image
#
###############################################################
NETWORK_WIDTH = 160
NETWORK_HEIGHT = 160
preprocessed_image = cv2.resize(src, (NETWORK_WIDTH, NETWORK_HEIGHT))
preprocessed_image = cv2.cvtColor(preprocessed_image,
cv2.COLOR_BGR2RGB)
preprocessed_image = self.OpenCV.whiten(preprocessed_image)
return preprocessed_image
def Compare(self, face1, face2):
###############################################################
#
# Determines whether two images are a match
#
###############################################################
if (len(face1) != len(face2)):
self.Helpers.logMessage(self.LogFile, "Facenet", "!ERROR!",
"Distance Missmatch")
return False
tdiff = 0
for index in range(0, len(face1)):
diff = np.square(face1[index] - face2[index])
tdiff += diff
if (tdiff < 1.3):
self.Helpers.logMessage(self.LogFile, "Facenet", "Classification",
"Calculated Match: " + str(tdiff))
return True, tdiff
else:
self.Helpers.logMessage(self.LogFile, "Facenet", "Classification",
"Calculated Mismatch: " + str(tdiff))
return False, tdiff
def addFrameBB(self, frame, x, y, w, h):
###############################################################
#
# Adds bounding box to the passed frame
#
###############################################################
cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)
return frame
def addFrameFeatures(self, frame, shape, x, y):
###############################################################
#
# Adds facial features to the passed frame
#
###############################################################
for (x, y) in shape:
cv2.circle(frame, (x, y), 1, (0, 255, 0), -1)
return frame
def connect(self): """ Connects to the USB camera. """ self.camera = OpenCV(self.Helpers.confs["Camera"]["Id"])
def connect(self):
""" Connects to the USB camera. """
self.camera = OpenCV(self.Helpers.confs["Camera"]["Id"])
self.Helpers.logger.info("Connected To Camera")
class Facenet():
""" ALL Detection System 2019 Facenet Class
Facenet helper functions for the ALL Detection System 2019 Facial Authentication Server project.
"""
def __init__(self, LogPath):
""" Initializes the Facenet class. """
# Class settings
self.Known = []
self.Helpers = Helpers()
self.LogFile = self.Helpers.setLogFile(self.Helpers.confs["Server"]["Logs"]+"/Facenet")
# OpenCV settings
self.OpenCV = OpenCV(self.Helpers)
# Dlib settings
self.Detector = dlib.get_frontal_face_detector()
self.Predictor = dlib.shape_predictor(self.Helpers.confs["Classifier"]["Dlib"])
def PreprocessKnown(self, ValidDir, Graph):
""" Preprocesses known images. """
for validFile in os.listdir(ValidDir):
if os.path.splitext(validFile)[1] in self.Helpers.confs["Classifier"]["ValidIType"]:
self.Known.append({"File": validFile, "Score": self.Infer(cv2.resize(cv2.imread(ValidDir+validFile), (640, 480)), Graph)})
self.Helpers.logMessage(self.LogFile,
"Facenet",
"STATUS",
str(len(self.Known)) + " known images found.")
def ProcessFrame(self, Frame):
""" Preprocesses frame. """
Known = []
Frame = cv2.resize(cv2.imdecode(Frame, cv2.IMREAD_UNCHANGED), (640, 480))
RawFrame = Frame.copy()
Gray = cv2.cvtColor(Frame, cv2.COLOR_BGR2GRAY)
Path = "Data/Captured/" + datetime.now().strftime("%Y-%m-%d") + "/" + datetime.now().strftime("%H") + "/"
FileName = datetime.now().strftime('%M-%S') + ".jpg"
FileNameGray = datetime.now().strftime('%M-%S') + "-Gray.jpg"
self.OpenCV.SaveFrame(Path + "/", FileName, Frame)
self.OpenCV.SaveFrame(Path + "/", FileNameGray, Gray)
return Frame
def LoadGraph(self):
""" Loads Facenet graph. """
with open(self.Helpers.confs["Classifier"]["Graph"], mode='rb') as f:
graphFile = f.read()
self.Helpers.logMessage(self.LogFile, "Facenet", "Status", "Loaded TASS Graph")
return graphFile
def Infer(self, face, graph):
""" Runs the image through Facenet. """
graph.LoadTensor(self.PreProcess(face).astype(np.float16), None)
output, userobj = graph.GetResult()
return output
def PreProcess(self, src):
""" Preprocesses an image. """
NETWORK_WIDTH = 160
NETWORK_HEIGHT = 160
preprocessed_image = cv2.resize(src, (NETWORK_WIDTH, NETWORK_HEIGHT))
preprocessed_image = cv2.cvtColor(preprocessed_image, cv2.COLOR_BGR2RGB)
preprocessed_image = self.OpenCV.whiten(preprocessed_image)
return preprocessed_image
def Compare(self, face1, face2):
""" Determines whether two images are a match. """
if (len(face1) != len(face2)):
self.Helpers.logMessage(self.LogFile, "Facenet", "!ERROR!", "Distance Missmatch")
return False
tdiff = 0
for index in range(0, len(face1)):
diff = np.square(face1[index] - face2[index])
tdiff += diff
if (tdiff < 1.3):
self.Helpers.logMessage(self.LogFile, "Facenet", "Classification", "Calculated Match: " + str(tdiff))
return True, tdiff
else:
self.Helpers.logMessage(self.LogFile, "Facenet", "Classification", "Calculated Mismatch: " + str(tdiff))
return False, tdiff