BIO

I am with Epic Games as a Director of R&D and Principal Researcher in AI & Computer Vision. Prior to that I was a scientist / researcher / engineer  at Snap Inc., DAQRI (acquired by Snap), Austrian Institute of Technology, a visiting research scholar at Stanford University, USA and a postdoctoral researcher at the Vienna University of Technology, Austria. I received Ph.D. in Artificial Intelligence and Bio-cybernetics, focusing on Computer Vision from the Czech Technical University in Prague, at the Center of Machine Perception. My research interests are driven by learning computers to understand what they see. I am keen on knowledge transfer from science to industry. I am a holder of the Microsoft Visual Computing Award 2011 given to the best young scientists in Visual Computing in Austria and the Best Scientific Paper Prize at the British Machine Vision Conference in '07. I served as a member on program committees of all major computer vision conferences (ICCV, CVPR, ECCV, ...), a reviewer of most important computer vision journals (PAMI, IJCV, ...).

EXPERTISE

Computer Vision and AI R&D:

• geometric deep learning, spatial AI, 3D, photogrammetry

• camera self-calibration, vision based localization, SLAM, SAM, SfM

• scene understanding, semantic image segmentation

• my entry point to computer vision by a self made mobile robot controlled by a camera [link]

PUBLICATIONS

check citations on google scholar

• Micusik B., Evangelidis G.: Ego-Motion Alignment from Face Detections for Collaborative Augmented Reality, arXiv 2020, [paper]

• Micusik B., Evangelidis G.: Renormalization for Initialization of Rolling Shutter Visual-Inertial Odometry,  International Journal of Computer Vision (IJCV), © 2021 Springer, [arXiv]

• Evangelidis G., Micusik B.: Revisiting Visual-Inertial Structure from Motion for Odometry and SLAM Initialization , Robotics and Automation Letters (RA-L), © 2021 IEEE, [arXiv, ieee]

• Micusik B., Wildenauer H.: Plane refined Structure from Motion, Scandinavian Conference on Image Analysis (SCIA), © 2017 Springer, [pdf]

• Micusik B., Wildenauer H.: Structure from Motion with Line Segments Under Relaxed Endpoint Constraints, International Journal of Computer Vision (IJCV), © 2016 Springer, DOI, [pdf]

• Micusik B., Wildenauer H.: Descriptor Free Visual Indoor Localization with Line Segments, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, USA, © 2015 IEEE [pdf]

• Micusik B., Wildenauer H.: Structure from Motion with Line Segments Under Relaxed Endpoint Constraints, International Conference on 3D Vision (3DV), Tokyo, Japan, © 2014 IEEE [pdf]

• Wildenauer H., Micusik B.: Closed form solution for radial distortion estimation from a single vanishing point, British Machine Vision Conference (BMVC), Bristol, UK, © 2013 [pdf]

• Micusik B., Wildenauer H.: Minimal solution for uncalibrated absolute pose problem with a known vanishing point, International Conference on 3D Vision (3DV), Seattle, USA, © 2013 IEEE [pdf]

• Hoedlmoser M., Micusik B., Kampel M.: Sparse Point Cloud Densification by Using Redundant Semantic Information, International Conference on 3D Vision (3DV), Seattle, USA, © 2013 IEEE [pdf]

• Hoedlmoser M., Micusik B., Pollefeys M., Liu M.-Y., Kampel M.: Model-Based Vehicle Pose Estimation and Tracking in Videos Using Random Forests, International Conference on 3D Vision (3DV), Seattle, USA, © 2013 IEEE [pdf]

• Hoedlmoser M., Micusik B.: Surface Layout Estimation Using Multiple Segmentation Methods and 3D Reasoning, Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA), Madeira, Portugal, 2013, © 2013 Springer

• Hoedlmoser M., Micusik B., Liu M.-Y., Pollefeys M., Kampel M.:Classification and Pose Estimation of Vehicles in Videos by 3D Modeling within Discrete-Continuous Optimization, International Conference on 3D Imaging, Modeling, Processing, Visualization and Transmission (3DIMPVT), Zurich, Switzerland, © 2012 IEEE. [pdf]

• Micusik B., Kosecka J., Singh G. :Semantic Parsing of Street Scenes of Video, International Journal f Robotics Research (IJRR), Volume 31, no. 4 ,pp 484-497, © 2012 SAGE Journals [pdf]

• Picus C., Pflugfelder R., Micusik B.:Auto-calibration of Non-overlapping Multi-camera CCTV Systems, Video Analytics for Business Intelligence, Studies in Computational Intelligence, Volume 409, pp 43-67, © 2012 Springer [pdf]

• Micusik B.: Trajectory reconstruction from non-overlapping surveillance cameras with relative depth ordering constraints, IEEE International Conference on Computer Vision (ICCV), Barcelona, Spain, © 2011 IEEE [pdf]

• Micusik B.: Relative pose problem for non-overlapping surveillance cameras with known gravity vector, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Colorado Springs, USA, © 2011 IEEE [pdf]

• Hödlmoser M., Micusik B., Kampel M.: Exploiting spatial consistency for object classification and pose estimation, International Conference on Image Processing (ICIP), Brussels, 2011

• Hödlmoser M., Micusik B., Kampel M.: Camera auto-calibration using Pedestrians and Zebra-Crossings, 11th IEEE Workshop on Visual Surveillance (VS2011) at International Conference on Computer Vision, Barcelona, Spain, 2011

• Picus C., Pflugfelder R., Micusik B.: Branch and bound optima search for tracking a single object and in a network of non-overlpapping cameras, 11th IEEE Workshop on Visual Surveillance (VS2011) at International Conference on Computer Vision, Barcelona, Spain, 2011 [pdf]

• Picus C., Micusik B., Pflugfeder R.: From single cameras to the camera network: an auto-calibration framework for surveillance, 32nd anuual symposium of the German Association for Pattern Recognition (DAGM), Darmstadt, 2010 [pdf]

• Micusik B., Kosecka J.: Multi-view Superpixel Stereo in Urban Environments, International Journal of Computer Vision (IJCV), Vol 89, Issue 1, 106-119, © 2010 Springer, DOI [pdf]

• Micusik B., Pajdla T.: Simultaneous surveillance camera calibration and foot-head homology estimation from human detections, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), San Francisco, USA, © 2010 IEEE [pdf]

• Micusik B., Pflugfelder R.:Localizing non-overlapping surveillance cameras under the L-infinity norm, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), San Francisco, USA, © 2010 IEEE [pdf]

• Donner R., Micusik B. Bischof H., Langs G.: Generalized sparse MRF appearance models, Journal on Image and Vision Computing (IVC), Elsevier, 28(6), 2010.

• Micusik B., Kosecka J.: Piecewise Planar City 3D Modeling from Street View Panoramic Sequences, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Miami, USA, © 2009 IEEE [pdf]

• Pflugfelder R., Micusik B.: Self-calibrating cameras in video surveillance, Chapter in the book Smart Cameras, Editor: A.N.Belbachir, Springer, 2009.

• Micusik B., Kosecka J.: Semantic segmentation of street scenes by superpixel co-occurrence and 3D geometry, IEEE Workshop on Video-Oriented Object and Event Classification (VOEC), held jointly with International Conf. on Computer Vision (ICCV), Japan, © 2009 IEEE [pdf]

• Kim Y.M., Theobalt Ch., Diebel J., Kosecka J., Micusik B., Thrun S.: Multi-view Image and ToF Sensor Fusion for Dense 3D Reconstruction. IEEE Workshop on 3-D Digital Imaging and Modeling (3DIM), held jointly with International Conf. on Computer Vision (ICCV), Japan, 2009.

• Micusik B., Wildenauer H. and Kosecka J.: Detection and Matching of Rectilinear Structures, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Anchorage, USA, © 2008 IEEE, [pdf]

• Murillo A.C., Kosecka J., Micusik, B., Sagues C., Guerrero J.J.: Weakly Supervised Labeling of Dominant Image Regions in Indoor Sequences, Vision in Action Workshop, held jointly with European Conference on Computer Vision (ECCV), Marseille, France, 2008.

• Micusik B., Wildenauer H. and Vincze M.: Towards Detection of Orthogonal Planes in Monocular Images of Indoor Environments, IEEE International Conference on Robotics and Automation (ICRA), Los Angeles, USA, © 2008 IEEE, [pdf]

• Wildenauer H., Micusik B. and Vincze M.: Efficient Texture Representation Using Multi-scale Regions, Asian Conference on Computer Vision (ACCV), Tokyo, Japan, 2007. [pdf]

• Donner R., Micusik B., Langs G. and Bischof H.: Sparse MRF Appearance Models for Fast Anatomical Structure Localisation, British Machine Vision Conference (BMVC), Warwick, UK, 2007. [pdf], - awarded Best Science Paper Prize -

• Micusik B. and Pajdla T.: Multi-label Image Segmentation via Max-sum Solver, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Minneapolis, USA, © 2007 IEEE, [pdf]

• Donner R., Micusik B., Langs G., Szumilas L., Peloschek P., Friedrich K. and Bischof H.: Object Localization Based on Markov Random Fields and Symmetry Interest Points , International Conf. on Medical Image Computing and Computer Assisted Intervention (MICCAI), Brisbane, Australia, Springer Verlag, 2007.

• Hanbury A., Micusik B. and Stöttinger J.: The MUSCLE Live Image Retrieval Evaluation Event , Proceedings of the Third MUSCLE/ImageCLEF Workshop on Image and Video Retrieval Evaluation, Hungary, 2007.

• Micusik B. and Hanbury A.: Template Patch Driven Segmentation, British Machine Vision Conference (BMVC), Edinburgh, UK, 2006. [pdf]

• Micusik B. and Hanbury A.: Automatic Image Segmentation by Positioning a Seed, European Conference on Computer Vision (ECCV), Graz, Austria, © 2006 Springer Verlag, [pdf]

• Micusik B. and Pajdla T.: Structure from Motion with Wide Circular Field of View Cameras, IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 28(7), © 2006 IEEE, [pdf]

• Szumilas L., Micusik B. and Hanbury A.: Texture Segmentation through Salient Texture Patches, Proceedings of the 11th Computer Vision Winter Workshop (CVWW),Telc, Czech Republic, 2006

• Micusik B. and Hanbury A.: Supervised Texture Detection in Images, Conference on Computer Analysis of Images and Patterns (CAIP), Paris, France, © 2005 Springer Verlag

• Micusik B. and Hanbury A.: Steerable Semi-automatic Segmentation of Textured Images, Scandinavian Conference on Image Analysis (SCIA), Joensuu, Finland, © 2005 Springer Verlag

• Micusik B. and Hanbury A.: Semi-automatic Segmentation of Textured Images, Proceedings of the 10th Computer Vision Winter Workshop (CVWW), Zell an der Pram, Austria, 2005

• Micusik B. and Pajdla T.: Autocalibration & 3D Reconstruction with Non-central Catadioptric Cameras, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Washington DC, USA, 2004, [pdf]

• Micusik B. and Pajdla T.: Estimation of Omnidirectional Camera Model from Epipolar Geometry, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Madison, USA, 2003, [pdf]

• Micusik B. and Pajdla T.: Omnidirectional Camera Model and Epipolar Geometry Estimation by RANSAC with Bucketing, Scandinavian Conference on Image Analysis (SCIA), Goteborg, Sweden, 2003, © Springer Verlag, [pdf]

• Micusik B. and Pajdla T.: Para-catadioptric Camera Auto-calibration from Epipolar Geometry, Asian Conference on Computer Vision (ACCV), Jeju Island, Korea, 2004 [pdf]

• Micusik B., Martinec D. and Pajdla T.: 3D Metric Reconstruction from Uncalibrated Omnidirectional Images, Asian Conference on Computer Vision (ACCV), Jeju Island, Korea, 2004 [pdf]

• Micusik B. and Pajdla T.: Using RANSAC for Omnidirectional Camera Model Fitting, Proceedings of the 8th Computer Vision Winter Workshop (CVWW), Valtice, Czech Republic, 2003, [pdf]

• Gachter S., Pajdla T. and Micusik B.: Mirror Design for an Omnidirectional Camera with a Space Variant Imager, IEEE Workshop on Omnidirectional Vision (OMNIVIS), Budapest, Hungary, 2001, [pdf]