Long name	Industrial Computer Vision
Approving CModule	IBA_BaET, BV2_BaMT, IBA_BaTIN
Responsible	Prof. Dr. Lothar Thieling Professor Fakultät IME
Level	Bachelor
Semester in the year	winter semester
Duration	Semester
Hours in self-study	78
ECTS	5
Professors	Prof. Dr. Lothar Thieling Professor Fakultät IME
Requirements	basic skills in signal processing basic skills in Java and/or C basic skills in analysis and linear algebra
Language	German
Separate final exam	Yes

image construction and access to image data
grey-level image and colour image
development environment
software design tools
compiler
linker
debugger
softwaretools for image processing and image analysis
softare-based access to image data and parameters
overview of the available ip-modules (moduls dor image processing and image analysis)
design and implementation of own ip-moduls
design of algorithmic chains based on ip-modules using visual programming

---

segmentation
histogram-based segmentation
histogram analysis
shading and its compensation
region-based segmentation
filling
split and merge
region growing
contour-based segmentation
contour tracking
hough-transformation

---

feature extraction
geometric features
basic features (area, perimeter, shape factor)
central moments
normalized central moments
polar distance
curvature
DFT of polar distance and curvature
color features (HSI)
texture features
co-occurrence matrix
haralick features

---

Klassifikation von Merkmalen
terms and concepts
feature vector, feature space, object classes
supervised / unsupervised classification
learning / not learning classification
typical methods
quader method
minimum distance
nearest neighbour
maximum likelihood
neuronale Netze
the artificial neuron as a simple classifier
operation
activation function
bias
training a neuron (gradient descent)
multi-layer-perceptron
operation
purposes of the layers
backpropagation training algorithm
development environment for creating and training neural networks
design and configuration of neural networks
training neural networks
verification tof rained networks
generating C-functions from trained networks

the presented methods for segmentation can be
named
described
delineated in terms of application areas
evaluated in terms of advantages and disadvanteges
problemspecific parameterized

---

the presented methods for feature extraction can be
named
described
delineated in terms of application areas
evaluated in terms of advantages and disadvanteges
problemspecific parameterized

---

the presented methods for scallsification can be
named
described
delineated in terms of application areas
evaluated in terms of advantages and disadvanteges
problemspecific parameterized

Type	Attendance (h/Wk.)
Lecture	2
Exercises (whole course)	0
Exercises (shared course)	0
Tutorial (voluntary)	0

none

purposeful handling of the tool chain for computer vision

---

deal with complex tasks in a small team

---

derive complex solutions that can be implemented using image processing and image analysis

Type	Attendance (h/Wk.)
Practical training	2
Tutorial (voluntary)	0

none