Viola and Jones Object Detection Framework in C++
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
bin First commit Mar 7, 2014
doc First commit Mar 7, 2014
lib First commit Mar 7, 2014
src First commit Mar 7, 2014
utils First commit Mar 7, 2014
README.md First commit Mar 7, 2014

README.md

Viola and Jones Object Detection Framework (C++)


This is an implementation of the [Viola and Jones Object Detection Framework][1] in C++. It has been particularly optimized for the face detection paradigm:
  • Horizontal flipping face sample images in training phase.
  • Rotating negative sample images in training phase.
  • 4 Haar features implemented.
  • Integral image for faster processing.
  • Post-normalization speedup.
  • Scale invariant (feature scaling).

Installation --------------
git clone https://github.com/alexdemartos/ViolaAndJones.git violaandjones
cd violaandjones/src
g++ -std=c++0x -lpng -O3 Feature.cpp WeakClassifier.cpp StrongClassifier.cpp CascadeClassifier.cpp main.cpp -o vandj `libpng-config --ldflags`
chmod +x vandj

Usage

  • Run ./vandj without arguments to see the options.
  • No classification models are provided for object/face detection, you must train your own.
  • No images nor training samples are provided.

Example

I've used this program to train a face detection algorithm. Here are some results: