Prejsť na obsah
Výskum na FIIT

Research in applications of computer vision and artificial intelligence methods in the domains of medical imaging and digital pathology for automatic processing of radiological and microscopic images for the purpose of qualitative and quantitative analysis.

Research of new techniques of large data visualization using VR and AR technologies and experimentation with interaction technique. Exploring and modeling human visual attention in different environments and tasks. Research into the explanability and interpretability of deep learning.

Research interests:

  • Application of novel neural network models for computer vision problems in domains of medical imaging and digital pathology.
  • Interpretability of neural network models to increase their credibility.
  • Image processing and analysis for pattern recognition, object detection and segmentation.
  • Scene understanding and semantic description - classification.
  • Novel interaction techniques in visualization, data visualization in virtual reality and augmented reality.
  • Computer graphics techniques for photorealistic real-time rendering, visualization, light-field capture and manipulation.
  • Visual attention modeling

 

Subgroup no. 1: Computer Vision and Medical Imaging Group

Keywords: Medical Imaging, Digital Pathology, Object Detection and Recognition

Computer vision is a computer science discipline with an ultimate goal to perceive, interpret and understand the images, signals, and other types of visual data. Our research is focused mainly on:

  • Visual object detection and object recognition: development of novel methods applying neural networks in traditional and modern problems of computer vision.
  • Medical image processing: CT and MRI radiological data analysis for the detection of anatomical anomalies and diseases, segmentation and 3D image registration.
  • Digital Pathology: Development of novel automatic methods supporting the diagnosis stated by human experts - histologists - by quantitative analysis of bioptical tissue slices.

 


Vanda Benešová
Researcher
e-mail: vanda_benesova[at]stuba.sk
website
Lukáš Graf
Researcher
e-mail: lukas.hudec[at]stuba.sk
website
Her research interest is focused at the fields of computer vision, image processing, signal processing and human-computer interaction. Research interest is oriented on computer vision, object segmentation detection and localization.

Marek Jakab
Researcher
e-mail: marek.jakab[at]stuba.sk
website
Matej Kompánek
Doctoral Student
e-mail: matej.kompanek[at]stuba.sk
website
Research interest covers computer vision and object detection. Research lies at computer vision, medical image processing, machine learning and deep learning in medical imaging.

Štefan Grivalský
Doctoral Student
e-mail: stefan.grivalsky[at]stuba.sk
website
Martin Tamajka
Doctoral Student
e-mail: martin.tamajka[at]stuba.sk
website
Research interest lies on computer vision, medical image processing, segmentation, ML and deep learning in medical imaging Research lies at computer vision, medical image processing, machine learning and deep learning in medical imaging.

 

Subgroup no. 2: HCI & Visualization & Graphics Group

Keywords: Visual Attention Prediction and Modeling, Augmented and Virtual Reality Applications, Big Data Visualization, Photorealistic rendering

The goal of computer graphics is to model objects in virtual environments and their effective and photorealistic rendering. Data visualisations aim to provide data understanding using visual representations.

Our research is focused mainly on:

  • Prediction of visual human attention: development of model of human visual attention, generation of visual saliency map
  • Exploring techniques for information visualisation in augmented and virtual reality
  • Creating software visualisations to enhance software engineering processes
  • Developing new interaction techniques for visualisations

 


Peter Kapec
Assistant Professor
e-mail: peter.kapec[at]stuba.sk
website
Researcher
e-mail: lukas.hudec[at]stuba.sk
website
Research interest lies at information and big data visualization, software visualization, graph visualization, visual analytics, novel interaction techniques in visualization, data visualization in virtual reality and augmented reality, source code analysis, software metrics and software representation via graph structures. Research interest is oriented on computer vision, object segmentation detection and localization.

Dominika Dolha
Doctoral Student
e-mail: dominika.dolha[at]stuba.sk
website
Miroslav Laco
Doctoral Student
e-mail: miroslav.laco[at]stuba.sk
website
His research lies at computer vision, visual attention modeling and egocentric video processing.

Lukáš Graf
Doctoral Student
e-mail: lukas.graf[at]stuba.sk
website
Konstiantin Rudenko
Doctoral Student
e-mail: konstiantin.rudenko[at]stuba.sk
website
His research lies at computer vision, visual attention modeling and egocentric video processing.

 

Selected recent publications

  1. POLATSEK, Patrik - WALDNER, Manuela - VIOLA, Ivan - KAPEC, Peter - BENEŠOVÁ, Vanda
    Exploring Visual Attention and Saliency Modeling for Task-Based Visual Analysis.
    Computers and Graphics. Vol. 72, (2018), s. 26-38. ISSN 0097-8493, Vol. 72, (2018), pp. 26-38. ISSN 0097-8493.
  2. FOGELTON, Andrej - BENEŠOVÁ, Vanda.
    Eye blink completeness detection. Comput. Vis. Image Underst., 2018.
  3. AYUGUN, R.S. - BENEŠOVÁ, Vanda
    Multimedia Retrieval that Works. 2018 IEEE Conf. Multimed. Inf. Process. Retr., 2018, pp. 63–68.
  4. MARTÁK, Lukáš Samuel - ŠAJGALÍK, Mário - BENEŠOVÁ, Vanda
    Polyphonic Note Transcription of Time-Domain Audio Signal with Deep WaveNet Architecture.
    2018 25th International Conference on Systems, Signals and Image Processing (IWSSIP), pp. 1–5, 2018.
  5. HUDEC, Ladislav - BENEŠOVÁ, Vanda
    Texture Similarity Evaluation via Siamese Convolutional Neural Network.
    2018 25th International Conference on Systems, Signals and Image Processing (IWSSIP), pp. 1-5, 2018.
  6. TAMAJKA, Martin - BENEŠOVÁ, Vanda
    Supervoxel Algorithm for Medical Image Processing.
    2017 IEEE Int. Conf. Power, Control. Signals Instrum. Eng, pp. 3121-3127, 2017.
  7. OLEŠOVÁ V. - BENEŠOVÁ, Vanda - POLATSEK, Patrik
    Visual Attention in Egocentric Field-of-view using RGB-D Data.
    Proceedings Ninth International Conference on Machine Vision (ICMV 2016), 18th November, 2016, Nice, France. 1. vyd : SPIE - The International Society for Optical Engineering, 2017.
  8. FOGELTON, Andrej - BENEŠOVÁ, Wanda
    Eye blink detection based on motion vectors analysis.
    Computer Vision and Image Understanding, 0, pp. 1–11. 2016.
    http://doi.org/10.1016/j.cviu.2016.03.011
  9. ILČÍKOVÁ, Ivana - BENEŠOVÁ, Wanda – POLEC, Jaroslav – CSÓKA, Tibor
    Texture aware image error concealment with fuzzy segmentation.
    Systems, Signals and Image Processing (IWSSIP), 2016 International Conference on. IEEE. 2016
    http://doi.org/10.1109/IWSSIP.2016.7502725
  10. POLATSEK, Patrik - BENEŠOVÁ, Vanda - PALETTA, Lucas - PERKO, R.
    Novelty-based Spatiotemporal Saliency Detection for Prediction of Gaze in Egocentric Video.
    IEEE Signal Processing Letters, 23(3), pp. 394–398. 2016.
    http://doi.org/10.1109/LSP.2016.2523339
  11. POLEC, Jaroslav - BENEŠOVÁ, Wanda - VARGIC, Radoslav - ILČÍKOVÁ, Ivana - CSÓKA, Tibor
    Texture Feature Extraction using an Orthogonal Transform of Arbitrarily Shaped Image Regions.
    Journal of Electronic Imaging, 25(6), 2016.
    http://doi.org/10.1117/1.JEI.25.6.061413
  12. TAMAJKA, Martin - BENEŠOVÁ, Vanda
    Automatic Brain Segmentation Method based on Super-voxels.
    International Conference on Systems, Signals and Image Processing IWSSIP 2016, Bratislava: IEEE. pp. 16–19, 2016.
  13. JAKAB, Marek - BENEŠOVÁ, Vanda - RAČEV, Marek
    3D Object Recognition based on Local Descriptors.
    IS&T/SPIE Electronic Imaging 2015. San Francisco, USA: SPIE. 2015.
  14. MARTIN-GUTIERREZ, Jorge - FABIANI, Pena - BENEŠOVÁ, Vanda - MENESES, Dolores Maria - MORA, Carlos E.
    Augmented Reality to Promote Collaborative and Autonomous Learning in Higher Education.
    Computers in Human Behavior, 51, pp.752–761, 2015.
    http://doi.org/10.1016/j.chb.2014.11.093
  15. POLATSEK, Patrik - BENEŠOVÁ, Vanda
    Bottom-up Saliency Model Generation using Superpixels.
    Proceedings of the 31st Spring Conference on Computer Graphics. ACM New York, NY, pp. 121–129, 2015.
    http://doi.org/10.1145/2788539.2788557
  16. UHLÍKOVÁ, Ivana - BENEŠOVÁ, Vanda - POLEC, Jaroslav - CSÓKA, Tibor
    Texture Aware Image Error Concealment.
    EUROCON 2015. Salamanca, Spain: IEEE Catalog Number: CFP15EUR-CDR. 2015.
  17. BENEŠOVÁ, Vanda - KOTTMAN, Michal
    Fast Superpixel Segmentation Using Morphological Processing.
    Proceedinks of the International Conference on Machine Vision and Machine Learning - MVML 2014. Avestia Publishing.
  18. ŠIKUDOVÁ, Elena - ČERNEKOVÁ, Zuzana - BENEŠOVÁ, Vanda - HALADOVÁ, Zuzana - KUČEROVÁ, Júlia
    Počítačové videnie Detekcia a rozpoznávanie objektov. Praha: Wikina, Praha. 2013.
    Retrieved from http://vgg.fiit.stuba.sk/wordpress/wp-uploads/2014/01/strukturovana_kniha_CD.pdf
  19. BENEŠOVÁ, Vanda - KOTTMAN, Michal - SIDLA, Oliver
    Hazardous Sign Detection for Safety Applications in Traffic Monitoring.
    Röning Juha & Casasent David P. (Eds.), Proceedings of SPIE, Intelligent Robots and Computer Vision XXIX: Algorithms and Techniques, Volume 8301. San Francisco, USA. 2012.
    http://doi.org/10.1117/12.905813
  20. SIDLA, Oliver - KOTTMAN, Michal - BENEŠOVÁ, Vanda
    Real-time Pose Invariant Logo and Pattern Detection.
    Proceedings of SPIE, Vol. 7878, p. 8, 2011
    Retrieved from http://vgg.fiit.stuba.sk/files/Real-time pose invariant logo and pattern detection.pdf
  21. SIDLA, Oliver - BRÄNDLE, Norbert - BENEŠOVÁ, Vanda - ROSNER, Marcin - LYPETSKYY, Yuriy
    Embedded Vision Challenges.
    A. N. Belbachir (Ed.), Smart Cameras. Boston, MA: Springer US. pp. 99–117, 2010.
    http://doi.org/10.1007/978-1-4419-0953-4
  22. SIDLA, Oliver - BRAENDLE, Norbert - BENEŠOVÁ, Vanda
    Towards Complex Visual Surveillance Algorithms on Smart Cameras
    Computer Vision Workshops (ICCV Workshops), 2009 IEEE 12th International Conference on Computer Vision. Kyoto: IEEE, 2009, pp. 847–853. Print ISBN: 978-1-4244-4442-7.
    http://doi.org/10.1109/ICCVW.2009.5457615
  23. KAPEC, Peter - BRNDIAROVÁ, G., GLOGER, M., MARÁK, J.
    Visual Analysis of Software Systems
    INES 2015 : 19th International conference on intelligent engineering systems. Bratislava, Slovakia. September 3-5, 2015. Danvers : IEEE, 2015, pp. 31-37, 2015.
  24. KAPEC, Peter - DRAHOŠ, Peter - ŠPERKA, Martin
    Trends in Information Visualization for Software Analysis and Development.
    Current Issues of Science and Research in the Global World: Proceedings of the International Conference on Current Issues of Science and Research in the Global World, Vienna, Austria; 27–28 May 2014. CRC Press, 2014. APA
  25. KAPEC, Peter - PAPRČKA, Michal - PAŽITNAJ, Adam - POLÁK, Vladimír
    Exploring 3D GPU-accelerated graph visualization with time-traveling virtual camera.
    JTACS - Journal of Theoretical and Applied Computer Science, Vol. 7, No. 2, pp 16-30, ISSN 2299-2634 (printed), 2300-5653 (online), 2013.
  26. GRZNÁR, Filip - KAPEC, Peter
    Visualizing Dynamics of Object Oriented Programs with Time Context.
    Proceedings of the 29th Spring Conference on Computer Graphics, pp. 65-72. ACM, 2013.

 

Important recent research results and research projects

  • KEGA 068UK-4/2011 Integrácia štúdia spracovania vizuálnej informácie a vytvorenie komplexných multimediálnych učebných materiálov.
  • VEGA:1/0625/14 Visual object class recognition in video sequences using a linkage of information derived by a semantic local segmentation and a global segmentation of visual saliency.

 

Industry collaboration

  • Siemens Healthcare - research in the area of medical imaging
  • QBSW - feasibility study of methods in video broadcasting (advertising detection)
  • Joanneum Research GmBH - research in the area of human visual attention

 

Academy collaboration

  • FEI STU in Bratislava - Prof. Polec
  • FMFI UK - Dr. Elena Šikudová, Dr. Zuzana Černeková
  • TU Wien in Bratislava - Assoc. Prof. Ivan Viola, Manuela Waldner

 

Infrastructure

Computer Vision and Computer Graphics Laboratory
Laboratory of Computer Vision and Graphics Working Group VGG - Vision & Graphics Group serves primarily for research activities and students have a opportunity to work on their team projects, diploma projects or bachelor projects in the laboratory.

Instrumentation:

  • Optical see-through AR glasses (Vuzix STAR 1200XLD)
  • Head-mounted display headset (Oculus Rift)
  • Interactive transparent projection foil (UGO)
  • 3D monitor 27'' NVIDIA 3D Vision
  • Projectors (3D Vision) 144Hz (BenQ XL2720Z)
  • Nvidia 3D Vision Glasses
  • Nikon D810 full-frame DSLR camera with 36.3 MP (2x) including the lens: Nikon 14-24mm, Nikon 24-70mm, Nikon AF-S 70-200mm
  • High-speed camera 340 fps 2 MP resolution Basler acA2000-340kc
  • Light field camera LYTRO ILLUM
  • Accurate spectrophotometer Konica Minolta CM-2500d
  • RGB-D data acquisition using Kinect

 


Oculus Rift + Leap Motion sensor

Eye tracker: Eye Tracking Glasses SMI ETG 2 Wireless