Navigation

Martin Hirzer

Martin Hirzer



E-Mail hirzer(at)icg.tugraz.at

Phone +43 316 873-5074

Office Room ID02050


Martin Hirzer View Martin Hirzer's profile on LinkedIn


Short CV

I received an MSc degree (Dipl.-Ing.) and a PhD degree (Dr.techn.) in Telematics from the Graz University of Technology in 2009 and 2014, respectively. Currently, I am a postdoctoral researcher at the Institute for Computer Graphics and Vision where I have been involved in the Person ReID, the OUTLIER, the KiwiVision Cloud, and the MobiTrick project. My research is focused on inter-camera person re-identification via (on-line) machine learning considering different feature representations.



Research


Relaxed Pairwise Learned Metric for Person Re-Identification, ECCV 2012

(Hirzer, Roth, Koestinger, Bischof)

Matching persons across non-overlapping cameras is a rather challenging task. Thus, successful methods often build on complex feature representations or sophisticated learners. A recent trend to tackle this problem is to use metric learning to find a suitable space for matching samples from different cameras. However, most of these approaches ignore the transition from one camera to the other. In this paper, we propose to learn a metric from pairs of samples from different cameras. In this way, even less sophisticated features describing color and texture information are sufficient for finally getting state-of-the-art classification results. Moreover, once the metric has been learned, only linear projections are necessary at search time, where a simple nearest neighbor classification is performed. The approach is demonstrated on three publicly available datasets of different complexity, where it can be seen that state-of-the-art results can be obtained at much lower computational costs.

The work was supported by the Austrian Science Foundation (FWF) project Advanced Learning for Tracking and Detection in Medical Workflow Analysis (I535-N23) and by the Austrian Research Promotion Agency (FFG) project SHARE in the IV2Splus program.

Full text (pdf), Poster (pdf)


Person Re-Identification by Efficient Impostor-based Metric Learning, AVSS 2012

(Hirzer, Roth, Bischof)

Recognizing persons over a system of disjunct cameras is a hard task for human operators and even harder for automated systems. In particular, realistic setups show difficulties such as different camera angles or different camera properties. Additionally, also the appearance of exactly the same person can change dramatically due to different views (e.g., frontal/back) of carried objects. In this paper, we mainly address the first problem by learning the transition from one camera to the other. This is realized by learning a Mahalanobis metric using pairs of labeled samples from different cameras. Building on the ideas of Large Margin Nearest Neighbor classification, we obtain a more efficient solution which additionally provides much better generalization properties. To demonstrate these benefits, we run experiments on three different publicly available datasets, showing state-of-the-art or even better results; however, on much lower computational efforts. This is in particular interesting since we use quite simple color and texture features, whereas other approaches build on rather complex image descriptions!

The work was supported by the Austrian Research Promotion Agency (FFG) within the project SHARE in the IV2Splus program.

Full text (pdf), Slides (pdf)


Person Re-Identification by Descriptive and Discriminative Classification, SCIA 2011

(Hirzer, Beleznai, Roth, Bischof)

Person re-identification, i.e., recognizing a single person across spatially disjoint cameras, is an important task in visual surveillance. Existing approaches either try to find a suitable description of the appearance or learn a discriminative model. Since these different representational strategies capture a large extent of complementary information we propose to combine both approaches. First, given a specific query, we rank all samples according to a feature-based similarity, where appearance is modeled by a set of region covariance descriptors. Next, a discriminative model is learned using boosting for feature selection, which provides a more specific classifier. The proposed approach is demonstrated on two datasets, where we show that the combination of a generic descriptive statistical model and a discriminatively learned feature-based model attains considerably better results than the individual models alone. In addition, we give a comparison to the state-of-the-art on a publicly available benchmark dataset.

This work has been supported by Siemens AG Österreich, Corporate Technology (CT T CEE), Austria, and the project SECRET (821690) under the Austrian Security Research Programme KIRAS.

Full text (pdf), Poster (pdf), PRID 2011 dataset



Publications

2014

  1. Multiple Model Fitting by Evolutionary Dynamics (bib)Michael Donoser, Martin Hirzer, and Dieter Schmalstieg In Proc. IEEE International Conference on Pattern Recognition (ICPR), 2014
  2. Mahalanobis Distance Learning for Person Re-Identification (bib)Peter M. Roth, Martin Hirzer, Martin Koestinger, Csaba Beleznai, and Horst Bischof In Person Re-Identification, pages 247-267, Springer, 2014
    (The original publication is available at www.springer.com)
  3. Combining Descriptive and Discriminative Information for Person Re-Identification (bib)Martin Hirzer Ph.D. Thesis, Graz University of Technology, Faculty of Computer Science, 2014

2013

  1. Pedestrian Detection, Tracking and Re-Identification for Search in Visual Surveillance Data (bib)Csaba Beleznai, Michael Rauter, Martin Hirzer, and Peter M. Roth In Proc. Conf. of the Hungarian Association for Image Processing and Pattern Recognition, 2013

2012

  1. Dense Appearance Modeling and Efficient Learning of Camera Transitions for Person Re-Identification (bib)Martin Hirzer, Csaba Beleznai, Martin Koestinger, Peter M. Roth, and Horst Bischof In Proc. IEEE International Conference on Image Processing (ICIP), 2012
  2. Person Re-Identification by Efficient Impostor-based Metric Learning (bib)Martin Hirzer, Peter M. Roth, and Horst Bischof In Proc. IEEE International Conference on Advanced Video and Signal-Based Surveillance (AVSS), 2012
  3. Relaxed Pairwise Learned Metric for Person Re-Identification (bib)Martin Hirzer, Peter M. Roth, Martin Koestinger, and Horst Bischof In Proc. European Conference on Computer Vision (ECCV), 2012
    (The original publication is available at www.springerlink.com)
  4. Large Scale Metric Learning from Equivalence Constraints (bib)Martin Koestinger, Martin Hirzer, Paul Wohlhart, Peter M. Roth, and Horst Bischof In Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2012

2011

  1. Person Re-Identification by Descriptive and Discriminative Classification (bib)Martin Hirzer, Csaba Beleznai, Peter M. Roth, and Horst Bischof In Proc. Scandinavian Conference on Image Analysis (SCIA), 2011
    (The original publication is available at www.springerlink.com)
  2. Multi-Cue Learning and Visualization of Unusual Events (bib)Rene Schuster, Samuel Schulter, Georg Poier, Martin Hirzer, Josef Birchbauer, Peter M. Roth, Horst Bischof, Martin Winter, and Peter Schallauer In Proc. 11th IEEE Workshop on Visual Surveillance (ICCV), 2011

2009

  1. Saliency Driven Total Variation Segmentation (bib)Michael Donoser, Martin Urschler, Martin Hirzer, and Horst Bischof In Proc. International Conference on Computer Vision, 2009
  2. An Automatic Hybrid Segmentation Approach for Aligned Face Portrait Images (bib)Martin Hirzer, Martin Urschler, Horst Bischof, and Josef A. Birchbauer In Proc. Workshop of the Austrian Association for Pattern Recognition, 2009

2008

  1. Marker Detection for Augmented Reality Applications (bib)Martin Hirzer Technical Report, Graz University of Technology, Inst. f. Computer Graphics and Vision, ICG-TR-08/05, 2008
  2. Segmentation of Face Images (bib)Martin Hirzer MSc. Thesis, Graz University of Technology, Faculty of Computer Science, 2008


Copyright 2010 ICG