Jan Dirk Wegner: Katalogdaten im Frühjahrssemester 2021

NameHerr Prof. Dr. Jan Dirk Wegner
(Professor Universität Zürich (UZH))
Adresse
I. f. Geodäsie u. Photogrammetrie
ETH Zürich, HIL D 44.2
Stefano-Franscini-Platz 5
8093 Zürich
SWITZERLAND
Telefon+41 44 633 68 08
E-Mailjan.wegner@geod.baug.ethz.ch
DepartementBau, Umwelt und Geomatik
BeziehungDozent

NummerTitelECTSUmfangDozierende
103-0274-AALImage Processing
Belegung ist NUR erlaubt für MSc Studierende, die diese Lerneinheit als Auflagenfach verfügt haben.

Alle anderen Studierenden (u.a. auch Mobilitätsstudierende, Doktorierende) können diese Lerneinheit NICHT belegen.
3 KP6RK. Schindler, J. D. Wegner
KurzbeschreibungThe objective of this lecture is to introduce the basic concepts of image formation and explain the basic methods of signal and image processing.
LernzielUnderstanding core methods and algorithms in image processing and computer vision and the underlying signal processing foundations.
Applying image processing algorithms to relevant problems in photogrammetry and remote sensing.
InhaltThe following topics will be covered in the course:
- Properties of digital images
- Signal processing/Sampling
- Image enhancement
- Image restoration: Spatial domain
- Image restoration: Fourier domain
- Color/Demosaicing
- Image compression
- Feature extraction
- Texture analysis
- Image segmentation
SkriptA script will be provided as PDF files on the lecture website.
LiteraturWe suggest the following textbooks for further reading:

Rafael C. Gonzalez, Richard E. Woods
Digital Image Processing
Prentice Hall International, 2008
ISBN: 013168728X

Rafael C. Gonzalez, Steven L. Eddins, Richard E. Woods:
Digital Image Processing Using MATLAB
Prentice Hall, 2003
ISBN: 0130085197
Voraussetzungen / BesonderesThe lecture is accompanied by programming assignments, that need to be completed in order to pass the course.
103-0274-01LBildverarbeitung3 KP2GK. Schindler, J. D. Wegner
KurzbeschreibungEinfuehrung in die grundlegenden Konzepte und Methoden der digitalen Bildverarbeitung.
Lernziel- Mathematische und statistischen Modellierung von Bildern
- Beherrschung grundlegender Methoden der digitalen Bild- und
Signalverarbeitung
- Kenntnis wichtiger Bildverarbeitungsoperationen und -algorithmen
- Auswahl und Anwendung der vorgestellten Rechenwerkzeuge
zum Lösen von Bildverarbeitungsaufgaben
- Verständnis der computer-gestützten Bildverarbeitung als Grundlage
von Fernerkundung, Photogrammetrie und computer vision
Inhalt• Digitale Bilder, Signalprozessierung, Abtastung
• Geometrische Transformationen
• Farbräume
• Filterung von Bildern, Bildrestauration und -verbesserung
• Morphologische Operationen
• Punkt- und Liniendetektion
• Ähnlichkeitsmasse und Matching von Bildern
• Bildsegmentierung
• Radiometrie von Satellitenbildern
851-0648-00LMachine Learning for Global Development Belegung eingeschränkt - Details anzeigen
Number of participants limited to 24

Prerequisite: Students on BSc or MSc level who have already successfully participated in a data science and programming course.
3 KP2GJ. D. Wegner, L. Hensgen, A. Rom
KurzbeschreibungIn this course students will learn theories of machine learning and its application to problems in the context of global development, with a focus on developing countries (e.g. predicting the risk of child labor or chances of a malaria outbreak). By the end of the course, students will be able to critically reflect upon linkages between technical innovations, culture and individual/societal needs.
LernzielThe objective of this course is to introduce students with a non-technical background to machine learning. Emphasis is on hands-on programming and implementation of basic machine learning concepts to demystify the subject, equip participants with all necessary insights and tools to develop their own solutions, and to come up with original ideas for problems related to the context of global development. Specific importance is placed upon the reconciliation of the predictions, which have been generated by automated processes, with the realities on the ground; hence the linkage between technical and social issues. This raises questions such as “In how far can we trust an algorithm?”, “Which factors are hard to measure and therefore not integrated in the algorithm but still crucial for the result, such as cultural and social influences?”. These questions will be discussed in the interdisciplinary group, equipping students with various perspectives on this crucial and very current debate.
InhaltThis course will give an introduction to machine learning with emphasis on global development. We will discuss topics like data preprocessing, feature extraction, clustering, regression, classification and take some first steps towards modern deep learning. The course will consist of 50% lectures and 50% hands-on programming in python, where students will directly implement learned theory as a software to help solving problems in global development.
Voraussetzungen / BesonderesThis course will give an introduction to machine learning with emphasis on applications in global development. It will consist of 50% lectures and 50% programming exercises (in python). Teaching assistants from the EcoVision Lab will help with all programming exercises without any needs for additional funding.

Students should bring their laptops to the exercises because we will program on laptops directly.

It is required that students enrolling in this course have successfully passed a course that deals with basic data science and are familiar with programming (preferably in Python).