Machine Vision

11B2060

Bachelor

German

5.0

only summer term

1 semester

The module ‘Machine Vision’ or ‘How Computers See and Understand’ examines how computers can perceive and interpret visual information from sensors such as cameras. Students first learn about concepts, techniques and algorithms of classical image processing (IP), i.e. the representation and processing of image data and the extraction of objects in binary images.

They then learn the basics of image analysis using artificial intelligence (AI) through various artificial neural networks (ANNs) for object recognition and classification. There is a particular focus on practical applications and evaluating in which use cases classical, algorithm-based IP is more appropriate and in which cases AI-based methods are more useful. In this way, students not only gain an understanding of the basic concepts of image processing, but also apply this knowledge specifically to real-world applications. The module promotes the development of skills for implementing machine vision applications in various scenarios.

1. Introduction
2. Image representation and storage
3. Image processing – point operations, filters (linear/non-linear), geometric transformations
4. Object extraction and representation in binary images
5. Fundamentals of artificial neural networks (ANN) – deep ANN, convolutional ANN, data sets
6. Features for object recognition (classical BV and ANN)
7. Object and image classification (classical and with ANN)

The total workload for the module is 150 hours (see also "ECTS credit points and grading").

• Project Report, written or
• oral exam

• Field work / Experimental work

Graded examination performance:

Project report (written): approx. 10–15 pages; explanation: approx. 20 minutes
Oral examination: see general section of the examination regulations
Ungraded examination performance:

Experimental work: Experiment: approx. 5 experiments in total

The module requires programming skills and mathematical knowledge (especially vector and matrix calculus).

Students who have successfully completed this module will have a basic understanding of image data representation, know how to extract information, and be familiar with basic image processing algorithms and the fundamentals of artificial neural networks for image recognition.

Students deepen their programming skills and learn about the steps involved in image processing, from pixel representation to the extraction of knowledge from images using selected algorithms and methods. 

After successfully completing this module, students will be able to implement image processing algorithms and solve simple classification tasks using artificial neural networks.

• Weinhardt, Markus

• Sch?ning, Julius
• Weinhardt, Markus