anurag14 / accessfaceid Goto Github PK

Face Detection, verification and recognition in Near real time on CPU brewing with all the SOTA all over

License: Apache License 2.0

Python 100.00%

livestream face-detection face-recognition real-time feature-extraction insightface vggface face-alignment deep-learning computer-vision

accessfaceid's People

Contributors

Stargazers

Watchers

Forkers

axmadjon andrewliang dwtcourses oxosec newteppi

accessfaceid's Issues

majority mode in utils/utils_insightface.py fails

statistics. Statistics Error: no unique mode; found 2 equally common values error

Face recognition needs metric learning enhancement

Recongition even after using insightface for one shot examples happens to have a lot of wrong labels being given.

text and fillbox overflow. make floating

make fill box and font size dynamic in size

Encode.py is not able to detect a face and crashes

For certain face pictures faced library is unable to detect a face in the image and crashes after it. Building a if-else / try catch for same is required.

Normalize all embedding and check what happens to recognition accuracy

majority mode in utils/utils_insightface.py face when no match

[Live Streaming] -----------------------------------------------------------
[FaceTracker Server] Found 1 faces!
[Base Server] Running inference...
[Live Streaming] -----------------------------------------------------------
[FaceTracker Server] Found 1 faces!
[Base Server] Running inference...
Traceback (most recent call last):
File "livestream/newmain.py", line 79, in
livestream.run()
File ".\modules\camera_server.py", line 40, in run
self.processs(frame)
File "livestream/newmain.py", line 53, in processs
_similar_face_name = self.face_db.who_is_this_face(_face_description)
File ".\modules\face_db.py", line 54, in who_is_this_face
who_is_this = compare_faces(self.faces_descriptions, self.faces_names, face_description, mode)
File ".\utils\utils_insightface.py", line 57, in compare_faces
name = face_distance_single_majority(known_face_encodings, names, face_query)
File ".\utils\utils_insightface.py", line 19, in face_distance_single_majority
name = max(set(name_list),key=name_list.count)
ValueError: max() arg is an empty sequence
[ WARN:0] terminating async callback

where to get vggface.h5?

hello, where to get vggface.h5? thankyou

Face alignment library uses inferior detection implicitly

class FaceAlignment:
    def __init__(self, landmarks_type, network_size=NetworkSize.LARGE,
                 device='cuda', flip_input=False, face_detector='sfd', verbose=False):
        self.device = device
        self.flip_input = flip_input
        self.landmarks_type = landmarks_type
        self.verbose = verbose

        network_size = int(network_size)

        if 'cuda' in device:
            torch.backends.cudnn.benchmark = True

        # Get the face detector
        face_detector_module = __import__('face_alignment.detection.' + face_detector,
                                          globals(), locals(), [face_detector], 0)
        self.face_detector = face_detector_module.FaceDetector(device=device, verbose=verbose)

Face recognition refers broken insightface implementation(include similarity also)

bug in modules/face_db.py generates incorret file

generates incorrect file link

Build a visualization for display of post aligned faces being fed to the network

It would be handy to have a visualization tool/script to check post alignment output.

Encoding doesn't use face alignment

Recognition doesn't work because encoding do not use face alignment

Tracker has no visualization for no faces in start. Also fails in first face visualization

(opencv-env) F:\RBS FaceReid\accessfaceid>python livestream/newmain.py
2019-05-21 12:13:52.029104: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
[LOG] Loading Encoded faces Database ...
[Camera Server] Camera is initializing ...
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 0 faces!
[FaceTracker Server] Found 1 faces!
[Base Server] Running inference...
Traceback (most recent call last):
File "livestream/newmain.py", line 74, in
demo.run()
File ".\modules\camera_server.py", line 40, in run
self.processs(frame)
File "livestream/newmain.py", line 51, in processs
_similar_face_name = self.face_db.get_similar_faces(_face_description)
File ".\modules\face_db.py", line 37, in get_similar_faces
print('[Face DB] Looking for similar faces in a DataBase of {} faces...'.format(len(self.faces)))
AttributeError: 'Model' object has no attribute 'faces'
[ WARN:1] terminating async callback

register/register.py fails after adding a unknown person

Traceback (most recent call last):
File "register/register.py", line 79, in
register.run()
File ".\modules\camera_server.py", line 40, in run
self.processs(frame)
File "register/register.py", line 58, in processs
_similar_face_name = self.face_db.who_is_this_face(_face_description, cores='single')
File ".\modules\face_db.py", line 54, in who_is_this_face
who_is_this = compare_faces(self.faces_descriptions, self.faces_names, face_description)
File ".\utils\utils_insightface.py", line 57, in compare_faces
face_value,name= face_distance(known_face_encodings, names, face_query)
File ".\utils\utils_insightface.py", line 11, in face_distance
face_dist_value = np.linalg.norm(face_encodings - face_query, axis=1)
ValueError: operands could not be broadcast together with shapes (16384,) (512,)
[ WARN:0] terminating async callback

Make face detection bounding boxes different color for known/unknown

Green bbox for --> known
Red bbox for --> unkwon