Eigenface #1 screenshot. Eigenface #2 screenshot.

Index of face-recognition

This sample code demonstrates a working face recognition engine based on Eigenfaces. To run this code you need:
  • Java SDK
  • JMF: Java Media Framework
  • JAMA: A Java Matrix Package.
  • Images to compare against (sixteen of them at least). If you don't have some, you can use "Babushka images".

Obtaining images from your camera

The eigenface engine requires a set of images to detect against. You can use this modified (or the ZIP file the required libraries. To use this capture program, make sure you have Java and JMF installed and do:

java DataSourceReader -monitor vfw://0

(The vfw://0 is the URL for your camera. You can find out which URL your camera has by running JMFRegistry and looking for the value of Locator).

Face Recognition

Make sure you have Jama-1.0.1.jar in your CLASSPATH, then just run:
(Screenshot of command prompt)
(The c:\konrad\images is where I stored the JPG images used to compare against.


Download the source zip file.


In the if you uncomment line 92,93,107, and 125, you can see the graphical output of the image during the face recognition process (and also save it).