Mathematical Optimization for Data Science Group, Saarland University

Welcome to the homepage of the

Mathematical Optimization for Data Science Group

Department of Mathematics and Computer Science, Saarland University, Germany

Image and Video Segmentation

Our Contribution:

In the context of motion based video segmentation, we contributed to the development of the first unsupervised object-level video segmentation method that works on natural videos [1]. The framework was originally introduced by Brox and Malik in [BM2010]. While the original work was designed to distinguish objects based on the long term analysis of their translational portion of the motion, we generalized the analysis to higher order motion models [2]. Due to computational reasons, the approaches mentioned so far generate a sparse labeling of the pixels in the video. In order to obtain a dense segmentation of the whole video, a robust interpolation is required [3], [4].
A slightly different approach was proposed in [5], where the starting point is a dense supervised labelling of the first frame of a video sequence, which is to be propagated to all other frames in the video in an unsupervised manner. The same problem was tackled by Nagaraja et al. [NSB2015] in a significantly more stable and flexible way.
For more results on motion segmentation including some example video, see also [LMB].

Publications on Motion Segmentation:

P. Ochs, J. Malik, T. Brox:
Segmentation of moving objects by long term video analysis.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(6):1187-1200, 2014.

P. Ochs, T. Brox:
Higher Order Motion Models and Spectral Clustering.
IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2012.

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.

P. Ochs, T. Brox:
Object segmentation in video: a hierarchical variational approach for turning point trajectories into dense regions.
IEEE International Conference on Computer Vision (ICCV), 2011.

N.S. Nagaraja, P. Ochs, K. Liu, T. Brox:
Hierarchy of Localized Random Forests for Video Annotation.
In A. Pinz, T. Pock, H. Bischof, F. Leberl (Eds.): Pattern Recognition (Proc. DAGM). Lecture Notes in Computer Science, Vol. 7476, Springer, 2012.