FaceReco is a face recognition system, which can learn and recognize faces from video. It is developed by Marko Linna at the University of Oulu in the Research Work on Information Processing course. The C++/Qt sources are available under the BSD 3-Clause license. This software was developed for research purposes and there is no other support available than this readme file.
Check this video to see how the system performs on the Honda/UCSD video database.
FaceReco utilizes 3rd party software Chehra for facial landmark detection and tracking. Faces are represented using Local Binary Patterns (LBP). For more detailed information about the system, read the following paper:
FaceReco runs on Windows 64-bit hosts (tested with Windows 7).
In order to compile FaceReco, the following dependencies must be installed.
The following steps are required in order to compile FaceReco with OpenCV and Chehra.
CHEHRA_ROOT
variable in FaceReco.pro to point to Chehra root folder.OPENCV_ROOT
variable in FaceReco.pro to point to OpenCV 2.4.9 root folder."opencv2/opencv.hpp"
.Copyright (c) 2015 Marko Linna.
Licensed under the BSD 3-Clause license.
Please site the following paper in your publications if FaceReco helps your research.
@inproceedings{linna2015online,
title={Online Face Recognition System Based on Local Binary Patterns and Facial Landmark Tracking},
author={Linna, Marko and Kannala, Juho and Rahtu, Esa},
booktitle={International Conference on Advanced Concepts for Intelligent Vision Systems},
pages={403--414},
year={2015},
organization={Springer}
}