cascades are binary classifiers (e.g: face or not) you cannot use them to recognize different ppl.
again, it cannot work, you will need to use something else
use a pretrained cnn like facenet or spherenet, that transforms face images into small (128/512) feature vectors, which can easily be stored into a db, and compared with simple L2 or cosine distance
cascades are binary classifiers (e.g: face or not) you cannot use them to recognize different ppl.
again, it cannot work, you will need to use something else
use a pretrained cnn like facenet or spherenet, that transforms face images into small (128/512) feature vectors, which can easily be stored into a db, (network ?)db, and compared with simple L2 or cosine distance
