def extractFaces1(filename):
global faceCnt
preprocessImage(filename)
face_locations = changeboxtype(
face_cascade.detectMultiScale(cv2.imread(tempFile), 1.7, 5))
print(face_locations)
# img = Image.open(tempFile)
# img.save(savePath+str(faceCnt)+'.jpg')
# faceCnt+=1
for face_location in face_locations:
cropImage(tempFile, face_location, True,
savePath + str(faceCnt) + '.jpg')
faceCnt += 1
def recognizeFacesInImage(filename):
global data, names
preprocessImage(filename)
img = cv2.imread(tempFile)
rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
boxes = face_recognition.face_locations(rgb)
encodings = face_recognition.face_encodings(rgb, boxes)
faces = []
if len(encodings) != 0:
faces = knn.predict(encodings)
faces = list(set(faces))
print(faces)
return faces
def recognizeFacesInImage(filename):
global data, names, totalFreq
preprocessImage(filename)
img = cv2.imread(tempFile)
rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
boxes = face_recognition.face_locations(rgb)
encodings = face_recognition.face_encodings(rgb, boxes)
faces = []
for encoding in encodings:
matches = face_recognition.compare_faces(data['encodings'], encoding)
freq = {}
percent = {}
for name in names:
freq[name] = 0
for i in range(0, len(matches)):
if (matches[i]):
freq[data['names'][i]] += 1
for name in names:
percent[name] = float('%.2f' %
(freq[name] / totalFreq[name] * 100))
curMax = 90
curName = 'Unknown'
for name in names:
if percent[name] >= curMax:
curMax = percent[name]
curName = name
if (curName in faces) == False:
faces.append(curName)
print(faces)
return faces
def recognizeSingleFace(filename):
global knn
global data, names
preprocessImage(filename)
img = cv2.imread(tempFile)
rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
boxes = face_recognition.face_locations(rgb)
encoding = face_recognition.face_encodings(rgb, boxes)
if len(encoding) == 0:
return 'Unknown'
encoding = encoding[0]
face = knn.predict([encoding])[0]
# prob = knn.predict_proba([encoding])[0]
# print(prob)
return face
def recognizeSingleFace(filename):
global data, names, totalFreq
preprocessImage(filename)
img = cv2.imread(tempFile)
rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
boxes = face_recognition.face_locations(rgb)
encoding = face_recognition.face_encodings(rgb, boxes)
if len(encoding) == 0:
return 'Unknown'
encoding = encoding[0]
matches = face_recognition.compare_faces(data['encodings'], encoding)
freq = {}
percent = {}
for name in names:
freq[name] = 0
for i in range(0, len(matches)):
if (matches[i]):
freq[data['names'][i]] += 1
for name in names:
percent[name] = float('%.2f' % (freq[name] / totalFreq[name] * 100))
curMax = 90
curName = 'Unknown'
for name in names:
if percent[name] >= curMax:
curMax = percent[name]
curName = name
face = curName
return face
def extractFaces2(filename):
global faceCnt
preprocessImage(filename)
# img = Image.open(tempFile)
# scale = 0.5
# width, height = img.size
# newsize = (int(width*scale), int(height*scale))
# img = img.resize(newsize)
# img.save(tempFile)
image = face_recognition.load_image_file(tempFile)
face_locations = face_recognition.face_locations(image)
print(face_locations)
# img = Image.open(tempFile)
# img.save(savePath+str(faceCnt)+'.jpg')
# faceCnt+=1
for face_location in face_locations:
cropImage(tempFile, face_location, True,
savePath + str(faceCnt) + '.jpg')
faceCnt += 1