/mtcnn-pytorch

Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks

Primary LanguageJupyter NotebookMIT LicenseMIT

MTCNN

pytorch implementation of inference stage of face detection algorithm described in
Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks.

Example

example of a face detection

How to use it

Just download the repository and then do this

from src import detect_faces
from PIL import Image

image = Image.open('image.jpg')
bounding_boxes, landmarks = detect_faces(image)

For examples see test_on_images.ipynb.

Requirements

  • pytorch 0.2
  • Pillow, numpy

Credit

This implementation is heavily inspired by: