Editor Reviews

Preprocessing Algorithms:

Face Recognition is like a cool tool that can be used to test out different ways of recognizing faces. You can try using Neural Networks with TensorFlow and Caffe, which are fancy methods. But before you get started, you gotta do some preprocessing on the images. Stuff like making them grayscale, cropping them, aligning the eyes, adjusting the brightness, and making them all fancy with filters. There are also some other tricks like Local Binary Pattern and Histogramm Equalization that you can use to make the images even better for feature extraction and classification.

TensorFlow:

You can totally use TensorFlow with SVM or KNN to extract features and classify stuff. If you want the Inception5h model, just grab it from Google Storage. And if you need the VGG Face Descriptor model, go ahead and download it from Dropbox. Now, here's the catch: for the Inception5h model to work, make sure you copy the file "tensorflow_inception_graph.pb" to "/sdcard/Pictures/facerecognition/data/TensorFlow". As for the VGG Face Descriptor model, you'll need both "VGG_FACE_deploy.prototxt" and "VGG_FACE.caffemodel" files.

Features

Face Recognition can totally be used as a cool testing framework for face recognition techniques.