Jump to content

EU Centre of ExcellenceISO 9001

ERCIMW3C MemberFraunhofer Project Center

Geometric Modelling and Computer Vision Laboratory

Head of department

Dmitrij Csetverikov
Dmitrij Csetverikov
Address: 1111 Budapest, Kende u. 13-17.
Mail address: 1111 Budapest, Kende u. 13-17.
Room number: K 315
Phone: +36 1 279 6161
Fax: +36 1 466 7503
E-mail: csetverikov.dmitrijEZT_TOROLJE_KI@EZT_TOROLJE_KIsztaki.mta.hu


A major goal of computer vision research at the Geometric Modelling and Computer Vision (GMCV) Laboratory is to build photo-realistic 3D-4D models based on multimodal sensor data. Sensor data of various physical origins, such as camera, laser scanner and CT images, are processed and combined to obtain rich and geometrically correct models of real-world objects and scenes. Algorithmic tools are developed, implemented and tested for the fusion of 2D and 3D data, feature detection in images and shapes, segmentation of images, curves, point sets and surfaces, classification and matching, surface and curve fitting, and morphing.

Main research areas

  • Reconstruction of static and dynamic, non-rigid objects
  • Description of deformable surfaces
  • Processing and registration of measured 3D data
  • Motion detection, tracking and analysis
  • Image and shape recognition
  • 3D-4D visualisation
  • Industrial and medical applications, preservation of cultural heritage

Major results

  • In the framework of a large national medical project, new 3D model of human knee has been developed based on the registration and segmentation of multimodal image data such CT, MR and X-ray. The model has been used to study knee motion patterns, support surgery and develop new prosthesis.
  • A 4D studio has been built which is unique in Central and Eastern Europe. The studio creates dynamic models of moving people. The GPU-based implementation enables real-time operation and visualisation.
  • A completely automatic 3D reconstruction system has been built that uses structured light and rotating table to create full-surface models for non-specular and moderately specular surfaces.
  • In cooperation with the international company Realeyes, a novel system for gaze tracking has been developed that uses a conventional web camera to determine gaze location on a monitor for a (nearly) static face.
  • In cooperation with the German company API and its Hungarian subsidiary, a novel software system has been developed for laser-based measurement of car tire grooves. A prototype of the equipment was awarded the Golden Trophy for innovation in the Services - Service Stations category at the Equip Auto Show 2011 in Paris.


Education, consulting and software development in the following areas:

  • 3D-4D reconstruction
  • Image processing
  • Motion and gaze tracking
  • Calibration of cameras and other sensors

More information:

alternatív szöveg

Important publications


Layer extraction in rodent retinal images acquired by Optical Coherence Tomography
Authors: Molnár, József; Csetverikov, Dmitrij; DeBuc, D. Cabrera; Gao, Wei; Somfai, G.M.
Date: 2012.


Perspective reconstruction and camera auto-calibration as rectangular polynomial eigenvalue problem
Authors: Pernek, Ákos; Hajder, Levente
Date: 2010.
Approximation-free running SVD and its application to motion detection
Authors: Csetverikov, Dmitrij; Axt, Attila
Date: 2010.


Dynamic texture detection based on motion analysis
Authors: Fazekas, Sandor; Amiaz, Tomer; Csetverikov, Dmitrij
Date: 2008.


Model based system for computer assisted knee surgery
Authors: Renner, Gábor; Szántó, György
Date: 2007.
Photorealistic 3D models of real-world objects
Authors: Jankó, Zsolt
Date: 2007.
Analysis and performance evaluation of optical flow features for dynamic texture recognition
Detecting regions of dynamic texture
Authors: Amiaz, Tomer; Fazekas, Alexandru; Csetverikov, Dmitrij; Kiryati, Nahum
Editor: Sgallari, F.; Murli, A.; Paragios, N.
Date: 2007.


Pre-registration of arbitrarily oriented 3D surfaces using a genetic algorithm