I’m amazed at how hard it is to find a simple and fast gamma correction code on OpenCV.
PS: I’m still very busy with master thesis.
Trying to train cascade classifiers for my HoG based detector takes up the entire CPU resources and the whole day to train a sample version. Thus I ended up diverting my attention to University’s PC instead, however compiling OpenCV and making opencv_traincascade run without error is not a simple task. Let’s see how it goes.
The HoG detector works though, just that it’s too slow without using cascade classifier. Continue reading
So, I just bought Sublime Text from http://www.sublimetext.com, which is supposed to be a very good cross platform text editor. I’ve been programming OpenCV applications using XCode, but rarely using editor or even compiling it from command line.
However, after tinkering a bit, I’ve managed to create a this built system for Sublime, which should allow you to compile and run OpenCV / C++ application from it. Just go to Tools > Build System > New Build System
My OpenCV was installed using HomeBrew, which has pkg-config information for opencv already created inside /usr/Local/Cellar/opencv/2.4.4/lib/pkgconfig/, thus you can utilize pkg-config to generate the include and library flags for g++ from it.
Test animated gif
I was frustrated while trying many ways to train my SVM and translating math equations into computer programs so that I could detect human in better ways. So I decided to do something random, like this:
A small 3D flock simulation developed by my friend and I. It’s an extension of normal 2D flock simulation, done in Processing.
Sorry guys, no update this week. Been busy finishing projects and preparing for thesis.