Robust Local Optical Flow Libary Documentation
V 1.2
|
This documentation contains a description of the RLOF library (available at http://www.nue.tu-berlin.de/menue/forschung/projekte/rlof/).
The RLOF library contains:
RLOF library supports the following methods published in articles :
The RLOFlib library is a feature tracking library. In this documentation we will use the definition of feature tracking. We use feature tracking in the following context. For a given set of point (feature points) defined at positions in an image at a time t, the Feature tracking method is able to track the positions of the feature points in the subsequent image at time t+1. In other words for each point of the feature set at time t the PLK and RLOF derived methods compute the motion at these positions from image at time t to the subsequent image at time t+1.
In order to increase the robustness of the proposed implementation and avoid dependencies we decided to no longer provide a GPU implementation. Instead we focus on CPU based parallelization techniques such as the Streaming SIMD Extension instruction set (SSE) and multi-threading with Threading Building Blocks( Intel TBB) or OpenMP. The RLOF library C++ is compiled as a shared library and provides matlab MEX wrapper functions. It supports Windows, Linux and Mac OS. The following table provides an overview of the available paralellezation for each version:
Windows (32-bit) | Windows (64-bit) | Linux (64-bit) | Mac OS(64-bit) | |
---|---|---|---|---|
SSE | yes | yes | yes | yes |
Intel TBB | yes | yes | no | no |
OpenMP | no | no | yes | no |
Read Section Sample to learn more how to use and install the library. To build an application using the RLOF library please consider the Terms of Use and read the Requirements.
The Matlab package contains Mex-functions for each OS. Call "help mex_RLOF" for a description of the arguments and see "sampleRLOF.m" for an example use of the function.
Their should be no additional requirements to your OS. The library was tested with:
The library was build using OpenCV 2.4.6. To enable the compatibility to another or no OpenCV version the we provide an alternative interface. Therefore disable the linking of OpenCV using the NOT_USE_OPENCV preprocessor flag.
08/01/2013 Release Version 1.1
17/11/2014 Release Version 1.2
The folder sample contains the RLOFsample.cpp file which demonstrates the usage of the RLOF library. The sample project has been created with Microsoft Visual Studio 2010.
To build the sample you have to set your OpenCV library files. This should be done by editing the include path and library path sections in the OpenCV property sheet files (OpenCV2_4_6Debug.vsprops, OpenCV2_4_6Release.vsprops, and the respective x64 property sheets). The sample does the following:
To run the sample please mention to announce the binary path "/RLOF/bin" e.g. by setting the environment PATH variable (e.g. PATH=D:/workspace/FIONA/Arbeit_Senst/publicProjects/RLOF/bin/ ) in your project debugging settings.
Tobias Senst: senst @nue .tu-b erli n.de
This file is the property of the author and Communication Systems Group, Technische Universitaet Berlin. All rights reserved. It may not be publicly disclosed, distributed, used, copied or modified without prior written authorization by a representative of Communication Systems Group, Technische Universitaet Berlin or the author. Any modified version of this document needs to contain this header.
You are free to use this software for whatever you like for non-commercial personal, non-commercial or academic usage. If you use this algorithm for a scientific publication, please cite one of the following paper:
@INPROCEEDINGS{Senst2014, AUTHOR = {Tobias Senst and Thilo Borgmann and Ivo Keller and Thomas Sikora}, TITLE = {Cross based Robust Local Optical Flow}, BOOKTITLE = {21th IEEE International Conference on Image Processing}, YEAR = {2014}, month = {okt}, pages = {1967--1971}, address = {Paris, France}, doi = {} } @INPROCEEDINGS{Senst2013, AUTHOR = {Tobias Senst and Jonas Geistert and Ivo Keller and Thomas Sikora}, TITLE = {Robust Local Optical Flow Estimation using Bilinear Equations for Sparse Motion Estimation}, BOOKTITLE = {20th IEEE International Conference on Image Processing}, YEAR = {2013}, month = {sep}, pages = {2499--2503}, address = {Melbourne, Australia}, doi = {10.1109/ICIP.2013.6738515} } @ARTICLE{Senst2012, AUTHOR = {Tobias Senst and Volker Eiselein and Thomas Sikora}, TITLE = {Robust Local Optical Flow for Feature Tracking}, JOURNAL = {IEEE Transactions on Circuits and Systems for Video Technology}, YEAR = {2012}, month={sep}, volume={22}, number={9}, pages={1377--1387}, doi={10.1109/TCSVT.2012.2202070} } @INPROCEEDINGS{Senst2011, AUTHOR = {Tobias Senst and Volker Eiselein and Ruben Heras Evangelio and Thomas Sikora} TITLE = {Robust Modified L2 Local Optical Flow Estimation and Feature Tracking}, BOOKTITLE = {IEEE Workshop on Motion and Video Computing}, YEAR = {2011}, MONTH = jan, EDITOR = {Eric Mortensen}, PAGES = {685--690}, ADDRESS = {Kona, USA}, DOI = {10.1109/WACV.2011.5711571}, }
It is not allowed to use any content of this package for any commercial use or any advertisement for upcoming commercial products. If you want to use any content for such a purpose please contact: Prof. Dr.-Ing. Thomas Sikora sikor. a@nu e.tu- berl in.de
Software provided by Technische Universitaet Berlin with this document is provided "AS IS" and any express of implied warranties including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose are disclaimed. In no event shall the author or contributors be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services, loss of use, data, or profits or business interruption) caused in any way out of the use of this software, even if advised of the possibility of such damage.