direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Background Subtraction

Gaussian mixture models have been extensively used and enhanced in the surveillance domain because of their ability to adaptively describe multimodal distributions in real-time with low memory requirements. Nevertheless, they
still often suffer from the problem of converging to poor solutions if the main mode stretches and thus over-dominates weaker distributions. Based on the results of the Split and Merge EM algorithm, we propose a solution to this problem. Therefore, we define an appropriate splitting operation and the  corresponding criterion for the selection of candidate modes, for the case of background subtraction. We further enhance the performance of the proposed model by incorporating information from a static objects detector algorithm.

Related Publications



  • Rubén Heras Evangelio, Michael Pätzold, Thomas Sikora
    Splitting Gaussians in Mixture Models
    9th IEEE International Conference on Advanced Video and Signal-Based Surveillance, Beijing, China, 18.09.2012 - 21.09.2012
    ISBN: 978-1-4673-2499-1
    Details BibTeX



We provide binaries of the SGMM-SOD library in order to help other researchers to compare their results or to use our work as a module for their research. The files contain a binary package for the Windows operating system and a minimal example on how to use the library. We have tried to keep the interface as simple as possible. Should you need a more flexible interface, or detect any bug in the library, please don't hesitate to contact me by mail.

The segmentation results of the provided library have been evaluated by using the ChangeDetection.net Video Database. To obtain more information on the dataset and on the achieved results, please visit www.changedetection.net.

Current version available:

  • Windows Visual Studio 2010 [x86]

To download the SGMM-SOD library, please fill out this form:


Terms of use

This library is provided for research purposes only and without express or implied warranty of any kind. Any commercial use is prohibited. We kindly request all authors using the library in their research work to cite the respective papers.

Zusatzinformationen / Extras