Introduction to Computer Vision (English)

Termin(e)
Lecture /w Excercise

Tuesday, 3 - 5pm; Monday 1-3pm; Friday 3-5pm G29-335; G29-K059

Prof. Dr.-Ing. Klaus Tönnies

Enrollment:

This course is limited to 36 students only. First, please duly enroll at the dedicated lecture page within the LSF (registration is open from the 25.02.2020-31.03.2020).
Since we expect many more applicants than places are available, students have to successfully pass an admission exam in order to be allowed to enroll in this course.

The admission exam is postponed until further notice as the university decided that there will be no lectures and exams until April 20, 2020. We will publish new exam details on this web site as soon as we know under which regulations the summer term will start. We will also notify all students who enrolled until the deadline (31.3.) by mail as to when and how the exam is going to take place.

Computer vision is derived from the term human vision (i.e. visual perception of the human being). Computer vision methods describe processes of visual perception algorithmically on different levels or replace them with equivalent processes, so that recorded camera images can be automatically interpreted or analyzed. Computer Vision may be differentiated into Early Vision and High Level Vision. Early Vision methods derive generic information such as depth or motion from images whereas High Level Vision focuses on the analysis of objects depicted in the images. Examples are object classification or object detection. The aim this lecture series is to present basic algorithms from Early Vision and High Level Vision, which are part of many methods for image-based analysis in industry and research.

Lecture contents:

  1. Introduction
  2. Early Vision – Noise Suppression, Scale Space and Segmentation
  3. Early Vision – Features from Images
  4. Early Vision – The reconstruction of Depth (introduction)
  5. Stereo Vision I – Epipolar Geometry
  6. Stereo Vision II – Correspondence Analysis
  7. Motion – Optical Flow and Structure from Motion
  8. High Level Vision – The detection of Objects
  9. Template Matching
  10. Variable Templates
  11. Part-Based Models
  12. Object Tracking
  13. Object Classification (Introduction)

Credits: 5

Language: English

Accompanying project:

To be allowed to take part in the examination, students have to successfully participate in an accompanying project. Details will be discussed within the first lecture.