IAM

OPENSOURCEFAN STUDYING
STUDYINGCOMPUTERSCIENCEANDMATH COMPUTERSCIENCE

Check out the latest superpixel benchmark — Superpixel Benchmark (2016) — and let me know your opinion! @david_stutz

ARTICLE

Paper “Superpixel Segmentation: An Evaluation”

After completing my bachelor thesis, I was encouraged to submit the results at the Young Researcher Forum of the German Conference on Pattern Recognition (GCPR) 2015. In this article, I want to share the paper as well as the corresponding poster.

Update. The source code corresponding to the evaluated superpixel algorithms and the used benchmark can be found on GitHub: davidstutz/superpixels-revisited.

Update. The LaTeX source of the paper and the poster can be found on GitHub: davidstutz/gcpr2015-superpixels.

The German Conference on Pattern Recognition (GCPR) is held in Aachen this year. After finishing my bachelor thesis at the Computer Vision Group of RWTH Aachen University (results and code available here), I was encouraged to submit my results at the Young Researcher Forum of GCPR 2015. Fortunately, the paper was accepted and I will be able to present my results as part of a poster session.

The submission was restricted to 6 pages (excluding references) in LNCS Springer format. The LaTeX template is available here. The paper as well as the poster can be downloaded below.

Abstract

In recent years, superpixel algorithms have become a standard tool in computer vision and many approaches have been proposed. However, different evaluation methodologies make direct comparison difficult. We address this shortcoming with a thorough and fair comparison of thirteen state-of-the-art superpixel algorithms. To include algorithms utilizing depth information we present results on both the Berkeley Segmentation Dataset [1] and the NYU Depth Dataset [2]. Based on qualitative and quantitative aspects, our work allows to guide algorithm selection by identifying important quality characteristics.

Paper (∼ 2.7MB)Poster (∼ 7.2MB)

The final publication is available at link.springer.com. Please cite as follows:

@incollection{Stutz:2015,
    title = {Superpixel Segmentation: An Evaluation},
    author = {Stutz, David},
    year = {2015},
    isbn = {978-3-319-24946-9},
    booktitle = {Pattern Recognition},
    volume = {9358},
    series = {Lecture Notes in Computer Science},
    editor = {Gall, Juergen and Gehler, Peter and Leibe, Bastian},
    doi = {10.1007/978-3-319-24947-6_46},
    publisher = {Springer International Publishing},
    pages = {555 -- 562},
}

References

  • [1] P. Arbeláez, M. Maire, C. Fowlkes, J. Malik. Contour detection and hierarchical image segmentation. Transactions on Pattern Analysis and Machine Intelligence, 33(5):898–916, 2011.
  • [2] N. Silberman, D. Hoiem, P. Kohli, R. Fergus. Indoor segmentation and support inference from RGBD images. European Conference on Computer Vision, pages 746–760, 2012.

What is your opinion on this article? Did you find it interesting or useful? Let me know your thoughts in the comments below or get in touch with me:

@david_stutz