Alexander Shekhovtsov


  • MSc. in applied mathematics from National Technical University of Ukraine “KPI”
  • Ph.D. in mathematical engineering from Czech Technical University in Prague

Research Interests

  • Discrete and continuous optimization with applications in computer vision.
  • Markov random fields (energy minimization in computer vision).
  • Machine learning.



  • A. Shekhovtsov, C. Reinbacher, G. Graber and T. Pock: Solving Dense Image Matching in Real-Time using Discrete-Continuous Optimization,CVWW 2016 [pdf][bib]
  • A. Shekhovtsov: Higher Order Maximum Persistency and Comparison Theorems(submitted to CVIU: SI on Inference & Learning of Graphical Models) 2014 [preprint].
  • A. Shekhovtsov: Exact and Partial Energy Minimization in Computer Vision, PhD Thesis, February 2013. [pdf][bib][slides].
  • A. Shekhovtsov, P. Kohli and C. Rother: Curvature Prior for MRF-Based Segmentation and Shape Inpainting, DAGM 2012. [pdf],[bib][slides].
  • A. Shekhovtsov, V. Hlavac: A Distributed Mincut/Maxflow Algorithm Combining Path Augmentation and Push-Relabel, IJCV 2012. [pdf],[bib][slides][code][preprint].
  • A. Shekhovtsov, V. Hlavac: On Partial Opimality by Auxiliary Submodular Problems, Control Systems and Computers, 2011(2). [pdf][bib].
  • A. Shekhovtsov, V. Hlavac: Joint Image GMM and Shading MAP Estimation, ICPR 2010. [pdf][bib], Technical Report [pdf]
  • A. Shekhovtsov, V. Hlavac: A Lower Bound by One-against-all Decomposition for Potts Model Energy Minimization, CVVW 2008.[pdf][bib][slides] 
  • P. Kohli, A. Shekhovtsov, C. Rother, V. Kolmogorov, P. Torr: On partial optimality in multi-label MRFsICML 2008: Proceedings of the 25th International Conference on Machine Learning. [pdf][bib][slides]
  • Shekhovtsov, J.D. Garcia-Arteaga, T. Werner: A Discrete Search Method for Multi-modal Non-Rigid Image RegistrationNORDIA 2008: Proceedings of the 2008 IEEE CVPR Workshop on Non-Rigid Shape Analysis and Deformable Image Alignment. [pdf][bib][slides]
  • T. Werner, A. Shekhovtsov: Unified Framework for Semiring-Based Arc Consistency and Relaxation LabelingComputer Vision Winter Workshop , St. Lambrecht, Austria, February 2007.
  • Shekhovtsov: Supermodular decomposition of structural labeling problem (in Russian). Control Systems and Computers #1, 2006, pp.39-48; Kiev, Ukraine. [pdf][bib]
  • B. Flach, D. Schlesinger, A. Shekhovtsov: A Higher Order MRF-Model for Stereo-ReconstructionPattern Recognition, LNCS vol. 3175, 2004, 440-446. [link]

Technical Reports

  • A. Shekhovtsov, P. Kohli, C. Rother: Curvature Prior for MRF-based Segmentation and Shape Inpainting, Research Report CTU--CMP--2011--11, Czech Technical University. [pdf][bib]
  • A. Shekhovtsov, V. Hlavac: A Distributed Mincut/Maxflow Algorithm Combining Path Augmentation and Push-Relabel, Research ReportK333--43/11, CTU--CMP--2011—03, Czech Technical University. [pdf][bib]
  • A. Shekhovtsov, V. Hlavac: Joint Image GMM and Shading MAP Estimation, Research Report K333–35/10, CTU–CMP–2010–03, Czech Technical University. 
  • A. Shekhovtsov, V. Kolmogorov, P.Kohli, V. Hlavac, C. Rother, P. Torr: LP-relaxation of binarized energy minimization, Research Report CTU--CMP--2007—27, Czech Technical University, update 2008. [pdf][bib]