[EDU] Self education in face recognition technologies
There are available two scripts:
- face_detect.py - simple script which uses OpenCV lib to detect face coming from Web Camera
- face_recognizer.py - simple script whicn can detect and "recognize" detected face coming from Web Camera
OpenCV 3.0.0.0 or never NumPy 1.11.0 or never skimage 0.12.3 scipy 0.17.1
To run thin you should install all dependencies
- How to install OpenCV you can read here, especial for Ubuntu (Linux) you can get instructions here
- How to install NumPy
- How to install skimage
- How to install scipy
Process of face recognizing is not an easy proceed. Currently exists a lot of different methods to do this. In current project was selected OpenCV as base lib for processing video stream and Local Binary Pattern Histogram method for recognize object.
Process of recognizing is divided by following steps:
- capture frame from web camera
- detect face
- recognizing:
- compare with knowing histograms
- if found similar face than show it
- if face is unknown than ask who is on face
To run application please make sure you have installed all packages and plugged in Web Camera
Running:
$ cd
<your_working_directory>
$ python face_recognizer.py -c lbpcascade_frontalface.xml
Face detection and recognize are different things.
Detection is only find and mark somehow object. Recognizing is more complicated thing.
face_detect.py
script is done only for showing same detection process without any additional.
This example shows how to detect faces but something similar can be done for any object. Only you need to find appropriate cascade file.