HomeContact
Computerz
Face detection in a life video stream
René Dohmen
August 13, 2010
1 min

For one of our software projects where investigating the possibilities of the Open Source image processing libraryopencv.

When installing openCV from the ubuntu packages it goes wrong with the python bindings for example. So I followed the instructions at http://opencv.willowgarage.com/wiki/InstallGuide%20%3A%20Debian to get a working version. I used the OpenCV from sourceforge stable. Both the listed SVN version fail to compile at my workstation...

So yes it’s a bitch; but it’s powerfull

Instead of the build instruction with make I had to replace make with cmake:

Then i did:

Then i did a make in the samples/c dir to build all the samples. All the samples build correctly.

This Is the facedetect algorithm; it detects one or more faces in a live video stream.

With some extra work I alos got it working on the MAC; the most easiest way is to download a Universal binary of OpenCV-2.0 with some patches; it will work ass an Mac OSX Framework straight from xCode.

I usedthe description at http://opencv.willowgarage.com/wiki/Mac_OS_X_OpenCV_Port to get it running; I downloaded the DMG for the framework from: http://vislab.cs.vt.edu/~vislab/wiki/images/4/44/OpenCV2.0.dmg

Have fun: this is it working with Snow Leopard:

It will detect all (human) faces with a reasonable accuracy:

image0
image0

image1
image1


Related Posts

Working with git submodules
October 15, 2023
3 min
© 2024, All Rights Reserved.
Powered by formatics

Quick Links

Advertise with usAbout UsContact Us

Social Media