State-of-the-art face detection and landmarks localization
Project description
Face Detector
Python package and Command Line Tool for state-of-the-art face detection and face landmark points localization. It gathers the techniques implemented in dlib and mtcnn, which can be easily switched between by setting a parameter in the FaceDetector class instantiation (dlib_5 is default if no technique is specified, use dlib_5 for dlib with 5 landmarks and dlib_68 for dlib with 68 landmarks).
How to Install:
First install C compiler: sudo apt-get install cmake, g++
Then this package: pip install face-detector
How to Use python package:
from face_detector import FaceDetector
img_addr = "path/to/image.[jpg/png/jpeg ...]"
# First parameter in FaceDetector constructor specifies face detection method (dlib: fl_5 or fl_68, mtcnn is default: mtcnn)
face_detector = FaceDetector()
faces = face_detector.get_faces(img_addr)
# Or to get the most prominent face in photo
main_face = face_detector.get_main_face(img_addr)
# Show image with bounding boxes and landmarks
import cv2
img = cv2.imread(img_addr)
for face in faces:
bb = face.bounding_box
landmarks = face.landmarks
cv2.rectangle(img, (int(bb.x), int(bb.y)), (int(bb.x + bb.w), int(bb.y+bb.h)), (0, 255, 0), 1)
for l in landmarks:
cv2.circle(img, (l.x, l.y), 2, (0,0,255))
cv2.imshow('img', img)
cv2.waitKey(0)
cv2.destroyAllWindows()
How to use Command Line Tool
foo@bar:~$ facedetector /home/foo/images/Yasser_Arafat.jpg
The previous command will display the image passed in arguments with a bounding box wrapping every face in the image. Fig. 1 shows the image displayed.
Fig. 1 Face detections as outputted by facedetector command line tool
Fig. 2 Face detections and landmarks as outputted by facedetector with -l (--landmarks) and -j (--only-main-face) options
foo@bar:~$ facedetector /home/foo/images/Yasser_Arafat.jpg -j -o /tmp/output.jpg -l
The previous command adds -j, -l and -o options, which capture the main face in the photo, adds landmark points and output the image with bounding boxes to the specified path, respectivelly. It also display the image in Fig. 2.
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 Distributions
Built Distribution
File details
Details for the file face_detector-0.4-py3-none-any.whl
.
File metadata
- Download URL: face_detector-0.4-py3-none-any.whl
- Upload date:
- Size: 9.1 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/2.0.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.36.1 CPython/3.7.4
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8cda48bd2bc73235f8d3c455619825fa2cb1d59914f8d1c5ab1f264b3f823c21 |
|
MD5 | 191f35eee59c6fb19dd9bae259093024 |
|
BLAKE2b-256 | aa64a66e87537fbf6ef8f6e3c7f600aa3c74ae50c91d612114cb062f07d2d248 |