Informationen für
Logo VIS

« Zurück

Korrespondenzprobleme in Computer Vision (in Englisch)
Typ: Hauptstudium / Vertiefungslinie
Semester: SS 2012
Umfang: 2V+2Ü
Studiengang: Diplom Informatik, Diplom Softwaretechnik
Dozent: Prof. Dr.-Ing. Andrés Bruhn
Beschreibung:

Correspondence problems are a central topic in computer vision. The basic task amounts to identifying and matching corresponding features in different images or views of the same scene. Typical examples for correspondence problems are (i) the estimation of motion information from consecutive frames of an image sequence (optic flow), (ii) the reconstruction of a 3-D scene from a stereo image pair, (iii) the registration of medical image data from different image acquisition devices (e.g. CT and MRT), and (iv) the analysis of the motion of fluid flows. The central part of this lecture is concerned with discussing the most important types of correspondence problems together with suitable models and algorithms for solving them.

This class requires undergraduate knowledge in mathematics ("e.g. Mathematik für Informatiker und Softwaretechniker"). Knowledge in image processing or computer vision is useful as well. Also basic knowledge in artificial intelligence can be helpful. The lectures will be given in English.

This class is particularly suited for those students who wish to pursue a diploma or master thesis in the CVIS group in the field of computer vision.

 

NEWS:

- LAST CHANCE: For the oral exam please register until Friday, June 22th personally at the VIS secretary. Students without registration cannot participate in the exam!

- Lecture 9 will be moved from 08.06.2012 to Thursday 21.06.2012.

- IMPORTANT: For the oral exam please register until Friday, June 8th at the LSF and additionally at the VIS secretary.

- Lecture 8 will be moved from 08.06.2012 to Thursday 14.06.2012.

- Lecture 5 will be moved from 18.05.2012 to Thursday 24.05.2012.

 

Vorlesungsfolien:

  1. Introduction, Overview (12.04.2012)
  2. Block Matching, Correlation Methods, Occlusion Detection, Interest Points, Feature-based Methods (19.04.2012)
  3. Optic Flow I: Local Differential Methods, Parametrisation Models (26.04.2012)
  4. Optic Flow II: Global Differential Methods, Horn and Schunck (03.05.2012)
  5. Optic Flow III: Advanced Constancy Assumptions, Large Motion (10.05.2012)
  6. Optic Flow IV: Advanced Data and Smoothness Terms (24.05.2012)
  7. Optic Flow V: High Accuracy Methods (24.05.2012)
  8. Optic Flow VI: Advanced Numerics (14.06.2012)
  9. Stereo Matching I: Projective and Epipolar Geometry (21.06.2012)
  10. Stereo Matching II: Estimation of the Fundamental Matrix (21.06.2012)
  11. Optic Flow + Stereo: Scene Flow Estimation (28.06.2012)
  12. Medical Image Registration: Mutual Information, Elastic and Curvature-Based Regularisation, Landmarks (05.07.2012)
  13. Particle Image Velocimetry: Div-Curl-Regularisation, Incompressible Navier Stokes Prior  (19.07.2012)
  14. Summary and Outlook (20.07.2012)

 

 

Übungen:

Übungstermine

  1. Übung 2   (20.04.2012)
  2. Übung 3   (27.04.2012)
  3. Übung 4   (04.05.2012)
  4. Übung 5   (11.05.2012)
  5. Übung 6   (25.05.2012)
  6. Übung 7   (Homework)
  7. Übung 8   (15.06.2012)
  8. Übung 9   (none)
  9. Übung 10  (22.06.2012)
  10. Übung 11  (29.06.2012)
  11. Übung 12  (06.07.2012)
  12. Übung 13  (13.07.2012)

 

Dateien für Programmieraufgaben

  1. Übung 2   (20.04.2012)
  2. Übung 4   (04.05.2012)
  3. Übung 6   (25.05.2012)
  4. Übung 8   (15.06.2012)
  5. Übung 11  (29.06.2012)
  6. Übung 13  (13.07.2012)

 

Musterlösungen

  1. Übung 2   (20.04.2012)
  2. Übung 3   (27.04.2012)
  3. Übung 4   (04.05.2012)
  4. Übung 5   (11.05.2012)
  5. Übung 6   (25.05.2012)
  6. Übung 7   (02.06.2012)
  7. Übung 8   (15.06.2012)
  8. Übung 10  (22.06.2012)
  9. Übung 12  (06.07.2012)


Bilder:
Internet-Seite:
Termine: Donnerstag, 9:45 - 11:15 Uhr in 0.363
Übungen: Freitag, 11:30 - 13:00 Uhr in 0.363
Tutor: Prof. Dr.-Ing. Andrés Bruhn

« Zurück