def trainFace(name, identity):
with open(tmpImgPath, 'rb') as f:
npImg = comm.bytes2NpImg(f.read())
if npImg is None:
return MSG['EMPTY']
npImg = scaledImg(npImg)
if identity not in FaceProcessor.persons:
FaceProcessor.persons[identity] = name
elif FaceProcessor.persons[identity] != name:
FaceProcessor.persons[identity] = name
align, net = FaceProcessor.align, FaceProcessor.net
bb = align.getLargestFaceBoundingBox(npImg)
if bb is None:
content = comm.npImg2Bytes(npImg)
msg = MSG['PROCEEDED']
msg['detail'] = 'noFace'
msg['content'] = content
return msg
# return {
# "type": "proceeded",
# 'detail': 'noFace',
# "content": content
# }
p1 = (bb.left(), bb.top())
p2 = (bb.right(), bb.bottom())
cv2.rectangle(npImg, p1, p2, (0, 255, 0), 2)
alignedFace = align.align(
96, npImg, bb, landmarkIndices=openface.AlignDlib.OUTER_EYES_AND_NOSE)
phash = str(imagehash.phash(Image.fromarray(alignedFace)))
if phash not in FaceProcessor.images:
rep = net.forward(alignedFace)
FaceProcessor.images[phash] = {'identity': identity, 'rep': rep}
content = comm.npImg2Bytes(npImg)
msg = MSG['PROCEEDED']
msg['detail'] = 'faceTrained'
msg['content'] = content
return msg
def rawFace():
npImg = getTmpImg()
if npImg is None:
return MSG['EMPTY']
npImg = comm.scaledImg(npImg)
content = 'data:image/jpg;base64,' + \
base64.b64encode(comm.npImg2Bytes(npImg))
msg = MSG['PROCEEDED']
msg['detail'] = 'rawImg'
msg['content'] = content
return msg
def rawFace():
with open(tmpImgPath, 'rb') as f:
npImg = comm.bytes2NpImg(f.read())
if npImg is None:
return MSG['EMPTY']
npImg = scaledImg(npImg)
content = comm.npImg2Bytes(npImg)
# msg = {
# "type": "proceeded",
# 'detail': 'rawImg',
# "content": content
# }
msg = MSG['PROCEEDED']
msg['detail'] = 'rawImg'
msg['content'] = content
return msg
def detectFace():
with open(tmpImgPath, 'rb') as f:
npImg = comm.bytes2NpImg(f.read())
if npImg is None:
return MSG['EMPTY']
npImg = scaledImg(npImg)
hasFace = FaceDetector.markAllFacesByDlib(npImg)
content = comm.npImg2Bytes(npImg)
# msg = {
# "type": "proceeded",
# 'detail': 'faceDetected' if hasFace else 'noFace',
# "content": content
# }
msg = MSG['PROCEEDED']
msg['detail'] = 'faceDetected' if hasFace else 'noFace'
msg['content'] = content
return msg
def recognizeFace():
# global recognizerInited
# if not recognizerInited:
# updateRecognizer()
# recognizerInited=True
le, clf, persons, align, net = \
FaceRecognizer.le, FaceRecognizer.clf, FaceProcessor.persons, FaceRecognizer.align, FaceRecognizer.net
if clf is None:
return MSG['EMPTY']
with open(tmpImgPath, 'rb') as f:
npImg = comm.bytes2NpImg(f.read())
if npImg is None:
return MSG['EMPTY']
npImg = scaledImg(npImg)
bb = align.getLargestFaceBoundingBox(npImg)
if bb is None:
content = comm.npImg2Bytes(npImg)
msg = MSG['PROCEEDED']
msg['detail'] = 'noFace'
msg['content'] = content
return msg
# return {
# "type": "proceeded",
# 'detail': 'noFace',
# "content": content
# }
alignedFace = align.align(
96, npImg, bb, landmarkIndices=openface.AlignDlib.OUTER_EYES_AND_NOSE)
rep = net.forward(alignedFace)
rep1 = rep.reshape(1, -1)
predictions = clf.predict_proba(rep).ravel()
maxI = np.argmax(predictions)
confidence = predictions[maxI]
identity, name = -1, 'unknown'
if confidence >= 0.7:
identity = le.inverse_transform(maxI)
name = persons[identity]
# phash = str(imagehash.phash(Image.fromarray(alignedFace)))
# if phash not in FaceProcessor.images:
# rep = net.forward(alignedFace)
# FaceProcessor.images[phash] = {
# 'identity' : identity,
# 'rep' : rep
# }
p1 = (bb.left(), bb.top())
p2 = (bb.right(), bb.bottom())
cv2.rectangle(npImg, p1, p2, (0, 255, 0), 2)
# for p in openface.AlignDlib.OUTER_EYES_AND_NOSE:
# cv2.circle(npImg, center=landmarks[p], radius=3, color=(0, 255, 0), thickness=-1)
cv2.putText(npImg,
str(identity) + ':' + name, (bb.left(), bb.top() - 5),
cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.6,
color=(0, 255, 0),
thickness=2)
content = comm.npImg2Bytes(npImg)
# msg = {
# "type": "proceeded",
# 'detail': 'faceRecognized',
# "content": content
# }
msg = MSG['PROCEEDED']
msg['detail'] = 'faceRecognized'
msg['content'] = content
return msg