Project Details

Algoritmy rozpoznávání obrazu

Project Period: 1. 1. 2006 – 31. 12. 2008

Project Type: grant

Code: GA201/06/1821

Agency: Czech Science Foundation

Program: Standardní projekty

English title
Algorithms of image recognition
Type
grant
Keywords

Linear classifiers, AdaBoost, FPGA, DSP, acceleration in hardware, feature
extraction, image processing, LRP, LRD

Abstract

The goal of the proposed project was research of algorithms of pattern
recognition in images, specifically classifiers and feature extraction methods.
Powerful implementation of classifiers and feature extractors are basis for
development of many computer vision methods and its applications that are
currently a "hot topic" and that are applicable in the industry, human-machine
communication, traffic applications, security, etc.

The main focus of
the proposed project was to develop such versions of the algorithms that would be
implementable on the image processing acceleration platforms being developed at
FIT BUT in Brno that are based on exploitation of programmable logic circuits
(FPGA) and digital signal processors (DSP). The next goal was to implement and
test such algorithms. It is expected that mainly linear statistical classifiers,
which are long-term research topic at FEE CTU in Prague, and other selected
algorithms.

Team members
Zemčík Pavel, prof. Dr. Ing., dr. h. c. (DCGM) – research leader
Publication Results

2008

  • BERAN, V.; HEROUT, A.; HRADIŠ, M.; ŘEZNÍČEK, I.; ZEMČÍK, P. Video Summarization at Brno University of Technology. ACM Multimedia. 2008 ACM International Conference on Multimedia Compilation E-Proceedings, TVS'08. New Yourk: Association for Computing Machinery, 2008. p. 31-34. ISBN: 978-1-60558-303-7. Detail
  • CHMELAŘ, P.; BERAN, V.; HEROUT, A.; HRADIŠ, M.; JURÁNEK, R.; LÁNÍK, A.; MLÍCH, J.; NAVRÁTIL, J.; ŘEZNÍČEK, I.; ŽÁK, P.; ZEMČÍK, P. Brno University of Technology at TRECVid 2008. Proceedings of TRECVID 2008. Gaithersburg: National Institute of Standards and Technology, 2008. p. 1-16. Detail
  • DOBEŠ, M.; ZEMČÍK, P. Zpracování obrazu a algoritmy v C#. Praha: BEN - technická literatura s.r.o., 2008. 144 s. ISBN: 978-80-7300-233-6. Detail
  • HEROUT, A.; BERAN, V.; HRADIŠ, M.; POTÚČEK, I.; ZEMČÍK, P.; CHMELAŘ, P. TRECVID 2007 by the Brno Group. Proceedings of TRECVID 2007. Gaithersburg: National Institute of Standards and Technology, 2008. p. 1-6. ISBN: 978-1-59593-780-3. Detail
  • HEROUT, A.; JOŠTH, R.; ZEMČÍK, P.; HRADIŠ, M. GP-GPU Implementation of the "Local Rank Differences" Image Feature. Proceedings of International Conference on Computer Vision and Graphics 2008. Lecture Notes in Computer Science. Heidelberg: Springer Verlag, 2008. p. 1-11. ISBN: 978-3-642-02344-6. Detail
  • HEROUT, A.; ZEMČÍK, P.; JURÁNEK, R.; HRADIŠ, M. Implementation of the "Local Rank Differences" Image Feature Using SIMD Instructions of CPU. Proceedings of Sixth Indian Conference on Computer Vision, Graphics and Image Processing. Bhubaneswar: IEEE Computer Society, 2008. p. 1-9. ISBN: 978-0-7695-3476-3. Detail
  • HRADIŠ, M. Framework for Research on Detection Classifiers. Proceedings of Spring Conference on Computer Graphics. Budmerice: Comenius University in Bratislava, 2008. p. 171-177. ISBN: 978-80-89186-30-3. Detail
  • HRADIŠ, M. ShotBoundary Detection with Simple Boosted Classifiers. Proceedings of the 12th Conference STUDENT EEICT 2008 Volume 4. Brno: Faculty of Electrical Engineering and Communication BUT, 2008. p. 266-270. ISBN: 80-214-3163-6. Detail
  • HRADIŠ, M.; HEROUT, A.; ZEMČÍK, P. Local Rank Patterns - Novel Features for Rapid Object Detection. Proceedings of International Conference on Computer Vision and Graphics 2008. Lecture Notes in Computer Science. Lecture Notes in Computer Science. Heidelberg: Springer Verlag, 2008. no. 12, p. 1-12. ISSN: 0302-9743. Detail
  • POLOK, L.; HEROUT, A.; ZEMČÍK, P.; HRADIŠ, M.; JURÁNEK, R.; JOŠTH, R. "Local Rank Differences" Image Feature Implemented on GPU. Proceedings of the 10th International Conference on Advanced Concepts for Intelligent Vision Systems. Lecture Notes In Computer Science; Vol. 5259. Berlin, Heidelberg: Springer Verlag, 2008. p. 170-181. ISBN: 978-3-540-88457-6. Detail

2007

  • BERAN, V.; HRADIŠ, M.; HEROUT, A.; CHMELAŘ, P.; POTÚČEK, I.; SUMEC, S.; ZEMČÍK, P. Video Summarization at Brno University of Technology. ACM Multimedia. Augsburg: Association for Computing Machinery, 2007. p. 1-4. ISBN: 978-1-59593-702-5. Detail
  • GRANÁT, J.; HEROUT, A.; HRADIŠ, M.; ZEMČÍK, P. Hardware Acceleration of AdaBoost Classifier. Workshop on Multimodal Interaction and Related Machine Learning Algorithms (MLMI). Brno: 2007. p. 1-12. Detail
  • ŠILHAVÁ, J.; BERAN, V.; CHMELAŘ, P.; HEROUT, A.; HRADIŠ, M.; JURÁNEK, R.; ZEMČÍK, P. Platform for Evaluation of Image Classifiers. Spring Conference on Computer Graphics. Budměřice: Comenius University in Bratislava, 2007. p. 103-109. ISBN: 978-80-223-2292-8. Detail
  • ŠILHAVÁ, J.; BERAN, V.; CHMELAŘ, P.; HEROUT, A.; HRADIŠ, M.; JURÁNEK, R.; ZEMČÍK, P. Testbench for Evaluation of Image Classifiers. Computer Graphics & Geometry, 2007, vol. 2007, no. 9, p. 31-47. ISSN: 1811-8992. Detail
  • ZEMČÍK, P.; HRADIŠ, M.; HEROUT, A. Local Rank Differences - Novel Features for Image. Poster SCCG 2007. Budmerice: 2007. p. 1-12. Detail
  • ZEMČÍK, P.; ŽÁDNÍK, M. Adaboost Engine. Proceedings of FPL 2007. Amsterdam: IEEE Computer Society, 2007. p. 1-5. ISBN: 1-4244-1060-6. Detail

2006

  • BERAN, V.; GRANÁT, J.; HEROUT, A.; ZEMČÍK, P. DSP and FPGA Based Raster Image Processing Architecture. Proceedings of the Digital Technologies 2007 Workshop. Žilina: Zilina University Publisher, 2006. p. 1-6. Detail
  • DRAHANSKÝ, M.; FUNK, W.; NÖTZEL, R. Liveness Detection based on Fine Movements of the Fingertip Surface. In IEEE - The West Point Workshop. West Point, New York: 2006. p. 42-47. ISBN: 1-4244-0130-5. Detail
  • ZEMČÍK, P.; HEROUT, A.; BERAN, V.; GRANÁT, J. Hardware Accelerated Image Analysis in FPGA. Proceedings of SCCG. Častá-Papiernička: Comenius University in Bratislava, 2006. p. 1-4. ISBN: 80-223-2175-3. Detail
Applied Results

2008

2007

Back to top