Mathematical Optimization Group, University of Tübingen

$logo.png$

Welcome to the homepage of the

Mathematical Optimization Group

Department of Mathematics, University of Tübingen, Germany

Robust interactive multi-label segmentation with an advanced edge detector

S. Müller, P. Ochs, J. Weickert and N. Graf

Abstract:
Recent advances on convex relaxation methods allow for a flexible formulation of many interactive multi-label segmentation methods. The building blocks are a likelihood specified for each pixel and each label, and a penalty for the boundary length of each segment. While many sophisticated likelihood estimations based on various statistical measures have been investigated, the boundary length is usually measured in a metric induced by simple image gradients. We show that complementing these methods with recent advances of edge detectors yields an immense quality improvement. A remarkable feature of the proposed method is the ability to correct some erroneous labels, when computer generated initial labels are considered. This allows us to improve state-of-the-art methods for motion segmentation in videos by 5-10% with respect to the F-measure (Dice score).



pdf	Bibtex	Publisher's link

Citation:
S. Müller, P. Ochs, J. Weickert, N. Graf:
Robust interactive multi-label segmentation with an advanced edge detector.

In B. Andres, B. Rosenhahn (Eds.): German Conference on Pattern Recognition (GCPR). Lecture Notes in Computer Science, Vol. 9796, 117-128, Springer, 2016.

Bibtex:
@inproceedings{MOWG16, title = {Robust interactive multi-label segmentation with an advanced edge detector}, author = {S. M{\"u}ller and P. Ochs and J. Weickert and N. Graf}, year = {2016}, editor = {B. Andres and B. Rosenhahn}, booktitle = {German Conference on Pattern Recognition (GCPR)}, series = {Lecture Notes in Computer Science}, publisher = {Springer}, volume = {9796}, pages = {117--128} }