Classroom visual sign in program
Project description
Face In
Sign in system with face recognition Based on code from https://github.com/serengil/deepface
Urs Utzinger, 2022
Requisits
pip3 install deepface
pip3 install camera-util
Create folder database and place images of known faces into them. Supports .jpg files.
Supports multiple images for same person, in database/First_Last/First_Last001.jpg
Configuration
detector_backend = "retinaface" # opencv*, ssd, mtcnn, dlib, retinaface
freeze_time = 5 # how will the results be displayed
frame_threshold = 5 # how many frames required to focus on a face, 5*
minimum_face_width = 180 # filter out small faces, depends on camera
model_name = "ArcFace" # VGG-Face*, Facenet, OpenFace, DeepFace, DeepID, Dlib, ArcFace or Ensemble
distance_metric = "cosine" # cosine*, euclidean, euclidean_l2
datafile = "./database/bme210_2022Spring.dat"
db_path = "./database" # Folder for .jpg files of known faces
Performance
Detector | Example exection time in ms | Works w glasses | Works w obstruction | Works w facemask |
---|---|---|---|---|
Retinaface | 172-233 | yes | yes | yes |
OpenCV | 40-56 | no | no | no |
SSD | 19-30 | yes | yes | no |
MTCNN | 873-1029 | yes | yes | no |
DLIB | 346-352 | yes | no | no |
Model | Example exection time in ms | Works w glasses | Works w obstruction | Works w facemask |
---|---|---|---|---|
ArcFace | ||||
VGG-Face | ||||
FaceNet | ||||
OpenFace | ||||
DeepID |
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
FaceIn-1.0.1.tar.gz
(17.5 kB
view hashes)
Built Distribution
FaceIn-1.0.1-py2.py3-none-any.whl
(21.2 kB
view hashes)
Close
Hashes for FaceIn-1.0.1-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 22087d4a99dec7e1c68d13d71282dd30f819d6b2287fc038560664c8b5b39d16 |
|
MD5 | d7b1228386b0ee0a45f3555ab9b4db0f |
|
BLAKE2b-256 | aa32eda43e23d4eaf2123de40d5f99c807cf564918d49b6a523ada0e6c49651d |