Dr. András Majdik

Ph.D. University of Cluj-Napoca

Email: majdik (at) ifi (dot) uzh (dot) ch

Office: Andreasstrasse 15, AND 2.20

Phone: +41 44 635 68 23

Short CV

Currently I am a postdoctoral researcher with Davide Scaramuzza at the Robotics and Perception Group, University of Zurich, Switzerland.

I received a PhD in 2011 form the Technical University of Cluj-Napoca (TU Cluj) in the field of Simultaneous Localization and Mapping of mobile robots. My PhD thesis was co-advised by Prof. Dr. Gheorghe Lazea from Robotics Research Group, TU Cluj and Prof. Dr. José A. Castellanos from Robotics, Perception and Real Time Group, University of Zaragoza. I graduated in 2008 from the Faculty of Automation & Computer Science, TU Cluj, Romania with an MSc in engineering.

I worked as a visiting researcher: from March to November 2010 at the Department of Computer Science and Systems Engineering, University of Zaragoza, Spain; from January to April 2011 and from July to September 2012 at the 3-D Vision and Mobile Robotics Research Group, Department of Control Engineering and Information Technology, Budapest University of Technology and Economics, Hungary.

Volunteering: I regularly review papers for the following conferences: IROS, ICRA, AQTR; I was a local organizing committee member at RSS�10 and AQTR�12.

Please contact me for an up to date CV.

Research interests

Visual SLAM, Autonomous micro air vehicles, Urban scene recognition


MAV urban localization from Google Street View data

We tackle the problem of globally localizing a camera-equipped micro aerial vehicle flying within urban environments for which a Google Street View image database exists.

To avoid the caveats of current image-search algorithms in case of severe viewpoint changes between the query and the database images, we propose to generate virtual views of the scene, which exploit the air-ground geometry of the system.

To limit the computational complexity of the algorithm, we rely on a histogram-voting scheme to select the best putative image correspondences.

The proposed approach is tested on a 2km image dataset captured with a small quadroctopter flying in the streets of Zurich.

The success of our approach shows that our new air-ground matching algorithm can robustly handle extreme changes in viewpoint, illumination, perceptual aliasing, and over-season variations, thus, outperforming conventional visual place-recognition approaches.


A. Majdik, Y. Albers-Schoenberg, D. Scaramuzza MAV Urban Localization from Google Street View Data, IROS'13, IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS'13, 2013. [ PDF ] [ PPT ] [ Video ] [ DATA ]

Adaptive appearance based loop-closing in heterogeneous environments

This work concerns the problem of detecting loop-closure situations whenever an autonomous vehicle returns to previously visited places in the navigation area. A novel probabilistic on-line weight updating algorithm is introduced for the bag-of-words description of the gathered images which takes into account both prior knowledge derived from an off-line learning stage and the accuracy of the decisions taken by the algorithm along time. An intuitive measure of the ability of a certain word to contribute to the detection of a correct loop-closure is presented. The proposed strategy is extensively tested in challenging, large-scale environments, characterized by the presence of buildings and urban furniture together with pedestrians and different types of vegetation.

Budapest City Center data set

The data set was collected in the historical city center of Budapest (Hungary). A hand-held stereo camera gathered lateral images along a 2 km long human trajectory throughout the environment. A total of 562 stereo pair images were recorded, from which a maximum of 200 loop-closure events were identified by the human operator. The experiment was carried out in a heterogeneous environment characterized by the presence of buildings and urban furniture together with pedestrians, cars and different types of vegetation. The covered area was composed of a multitude of different urban scenes: park, narrow and wide streets with car traffic, pedestrian streets and public squares.

In case you would like to get a copy of the data set, please don�t hesitate to contact me.


A. Majdik, D. G�lvez-L�pez, G. Lazea, J.A. Castellanos, Adaptive Appearance Based Loop-Closing in Heterogeneous Environments, IROS'11, IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS'11, 2011. [ PDF ] [ Video ]
A. Majdik, D. Lupea, G. Lazea, L. Vajta, Simultaneous Localization and Mapping Using Adaptive Appearance Based Loop-Closing Detection, AQTR'12, IEEE International Conference on Automation, Quality and Testing, Robotics, AQTR'12, 2012. [ PDF ]

Contact me!


Robotics and Perception Group

Department of Informatics

University of Zurich

Andreasstrasse 15

8050 Zurich



Dr. András Majdik

Room: AND 2.20

Email: majdik@ifi.uzh.ch

Office: +4144 635 6823