This is a simple face detection example of using machine learning algorithms to search faces within a picture. It originates from Shantnu Tiwari's tutorial -- Face Recognition with Python, in Under 25 Lines of Code and Face Detection in Python Using a Webcam. It uses OpenCV cascade to break the problem of detecting faces into multiple stages. The algorithm starts at the top left of a picture and moves down across small blocks of data. During the moves, a series of coarse-to-fine quick tests are carried out on each block. And it will only detect a face if all stages pass.
See the github repository for examples of its usage: https://github.com/simonzhaoms/facedetect
-
To install and demostrate the algorithm:
$ pip3 install mlhub $ ml install facedetect $ ml configure facedetect $ ml demo facedetect
To detect faces:
-
From a local image file:
$ ml score facedetect ~/.mlhub/facedetect/images/abba.png -
From an image on the web:
$ ml score facedetect https://github.com/opencv/opencv/raw/master/samples/data/lena.jpg -
From your camera:
$ ml live facedetect
Sometimes the algorithm will fail to detect real faces, then you need to fine-tune the parameters to get the ideal results:
$ ml score facedetect https://github.com/ageitgey/face_recognition/raw/master/tests/test_images/obama.jpg
$ ml score facedetect https://github.com/ageitgey/face_recognition/raw/master/tests/test_images/obama.jpg --scaleFactor 1.3
$ ml live facedetect --scaleFactor 1.3 --minSize 7 --minNeighbors 40