Sections
 
You are here: Home › Members › Michael Donoser › AK Computer Vision KU › AK Computer Vision KU 2011/2012

AK Computer Vision KU 2011/2012

Diese KU begleitet die zugehörige Vorlesung AK Computer Vision. Ziel ist die detaillierte Recherche und Aufarbeitung einer Methode in einem, in der Vorlesung vorgestellten, wissenschaftlichen Themenbereich und die Implementierung eines Prototyps in Matlab. Zusätzlich muss ein wissenschaftliches Dokument verfasst werden, in dem die Methode detailliert beschrieben, und eine umfangreiche experimentelle Analyse der Funktionalität durchgeführt werden muss.

News:

  • This lecture (for WS 2011/12) started on the 25th october 2011

Scheduled dates:

  • 25. October 2011, 08:30-10:00, seminar room ICG IE02082

    Description lecture AKCV-KU 2010/2011

    Group assignment

    Topic assignment

  • 20. November 2011, 23:59: Submission of project proposal

    Research of state-of-the-art in the selected topic field (What is SOTA? Who has done work in the field? What are the most important references?)

    What method will get implemented?

    Is there code publicly available?

    How can the method be evaluated?

    Writing of a proposal (a few pages)

  • 21th November - 2th December 2011: Groupwise meeting

    Disucssion of proposal in office Donoser IE02042

    Appointments will be fixed per Email

  • 22th January 2012, 23:59 Submission Matlab Implementation and report as PDF oder Doc

    (1) Submission Matlab Code as ZIP per E-Mail an michael.donoser@tugraz.at

    (2) Submission Report as PDF oder DOC per E-Mail an michael.donoser@tugraz.at

    Report has to include: Description of (a) research topic in general (b) selected method in detail (c) implementation details (d) exhaustive experimental evaluation - Approximately 10 pages per group. Prefered style: ICG-Latex Style

  • 23th January 10:00-13:00: Final presentation in seminar room ICG IE02082

    Presentation time: 15 minutes per group plus 10 minutes discussion - Summary of work with main focus on experiments (Presentation as PPT on USB or usage of own noteboook). Participation is obligatory and an important part of the final grade.

Slides

  • Download Slides from preliminary discussion.

Topic Assignments

  • Multi-Label Image Segmentation: Hutter+Kampl
  • Objectness: Kendler + Oberweger
  • Convex Graph Matching: Öttl + Hofer
  • Descriptor BRIEF: Herzog + Heran
  • Zero Shot Learning: Staber + Wiesmaier
  • Patch Match Image Labelling: Schulter + Innerhofer
  • Semi-Supervised Learning: Wirtl + Freiberger
Document Actions
[Powered by Plone]