Facial Recognition Using Principle Component Analysis (PCA), Support Vector Machine (SVM), and OpenCV

GENERAL NOTES:

This git repo includes some training photos, 40 photos for 3 people for a total of 120 photos to train with. There is a sample photo that is used as a test to test the results of the training.

VERY IMPORTANT:

There is a file, haarcascade_frontalface_default.xml, that must stay within the directory of the program, it is necessary for OPENCV python package.

UPDATE 2018, not sure anymore with opencv3 ^

DEPENDENCIES:

1. OPENCV
2. Numpy
3. SciKit-Image AND SKlearn
4. Matplotlib
5. Scipy
6. PYTHON 3

INSTALLING OPENCV

Getting OPENCV installed can be really annoying. I strongly recommend using Homebrew to install it. If you dont have Homebrew installed, you can get it here: https://github.com/Homebrew/install

After installing homebrew, run:

brew install opencv

This is where it gets really tricky depending how you have your python enviornments set up: IE: having python2 and python3 and may even anaconda too. I can explain it, but this link most certainly does a better job at it: https://www.learnopencv.com/install-opencv3-on-macos/

If you need really help, you can send flag an issue and i'll get back relatively quickly.

Results

If you use the photos I provided, you'll get something around 80% success rate, which is pretty good for a standard machine learning program but obviously not the latest of the latest facial recog tools.

HOPE YOU LIKE IT.