AkiraXD0712/Face_Recognition

Real-time face recognition by using keras(Tensorflow as backend), can be applied on Raspberry Pi3

Face_Recognition

Real-time face recognition by using Keras(Tensorflow as backend), can be applied on Raspberry Pi3.

Installation

We need to download python3.5 for using tensorflow on Windows, and we also need to download some python packages such as: numpy, opencv, scikit-learn, scipy, h5py, tensorflow, keras... For Tensorflow-GPU, we need to download CUDA Toolkit8.0.

For Windows

All packages used can be download on http://www.lfd.uci.edu/~gohlke/pythonlibs/.
Use python -m pip install <packages.xml> to install, run as root.

For Linux

sudo pip3 install numpy opencv-python scipy scikit-learn h5py

Install Tensorflow and Keras

For Tensorflow, see https://www.tensorflow.org/install/install_linux.
For Keras, use sudo pip3 install keras.

For Raspbian

sudo pip3 install numpy opencv-python scipy scikit-learn h5py

Install Tensorflow and Keras

For Tensorflow, see https://github.com/samjabrahams/tensorflow-on-raspberry-pi.
For Keras, use sudo pip3 install keras.

Creating your own data set

python training.py --input_dir=<> --model_dir=<> --epoch=<> --batch_size=<> --data_augmentation=<>