This material is copyrighted (c) 2014-2016 by Dr.-Ing. Stefan Steidl, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany. It is based on the former lectures Mustererkennung I & II by Prof. Dr.-Ing. Heinrich Niemann and the replacing lectures by Prof. Dr.-Ing. Joachim Hornegger and Elli Angelopoulou, PhD. Use without prior written permission of the authors is not permitted!

These slides are part of a presentation. The presentation is not complete without the verbal explanations given during the lecture.

The pdf files are password protected. You will get the password during the first lecture.

Three different versions of the slides are available:

1. The 'screen' version are the slides presented during the lectures. They may contain embedded video and audio files leading to large file sizes. Furthermore, they use extensively overlays, i.e. slides are displayed incrementally, resulting in a sequence of pdf pages instead of a single one just showing the final outcome. Hence, this version is not suited for printing. The slides are best viewed with a software that is able to show overlays without transition effects (e.g. GoodReader for iPad).
2. 'Print (A5)' shows two slides per page and is intended for printing booklets in the DIN A5 format. Let the printer driver rearrange the pages and use duplex printing. The printed versions do not contain video and audio files and most important, they do not have overlays. The printed versions include some slides that are especially designed for printing: for example, in a presentation, the difference between two similar images can be best viewed if both images are overlaid. As this is not possible in a print-out, both images are printed next to each other instead.
3. 'Print (A6)' shows four slides per page. Besides the different arrangement of slides per page, the slides are identical to the other print version. This is the standard option for printing if you want to file your print-outs in a folder.

We will update the slides during the semester as needed in order to improve the quality of our slides.
Please make sure that you have the latest version of our slides in preparation for the exam!

Additional material

Python code examples are shown during the lecture.
The algorithms are demonstrated using IPython Notebook.

Getting familiar with the algorithms is essential and will be part of the oral exams. This is true for both the 5.0 ECTS and the 7.5 ECTS exams. Please make sure that you have the latest version of the files as they might be updated.

• images:
  • faces.zip
  • images.zip

• videos:
  • colorplane.mp4
  • covmat.mp4
  • gaussian.mp4

• IPython Notebook files:
  • GettingStarted.ipynb (tutorial)
  • PlottingFunctions.ipynb (tutorial)
  • FourierSeries.ipynb (25.10.2017, requires ffmpeg to be installed)
  • SawtoothDFT.ipynb (14.11.2017)
  • Aliasing.ipynb (requires a recent version of IPython and a browser that supports html5 audio playback for the wav format)
  • WorkingWithImages.ipynb (tutorial)
  • HistogramEqualization.ipynb (14.11.2017)
  • ImageThresholding.ipynb (21.11.2017)
  • Variance.ipynb (21.11.2017)
  • UniformPDF.ipynb (28.11.2017)
  • BoxFilterFrequencyResponse.ipynb (28.11.2017)
  • FilteringImages.ipynb (28.11.2017)
  • GradientVectorField.ipynb (28.11.2017)
  • DetectingEdges.ipynb (28.11.2017)
  • RecursiveFilter.ipynb (29.11.2017)
  • MorphologicalOperations.ipynb (05.12.2017)
  • CosineImages.ipynb (13.12.2017)
  • WindowFunctions.ipynb (19.12.2017)
  • 2D-DFT.ipynb (13.12.2017)
  • 2D-DFT-Examples.ipynb (13.12.2017)
  • WHT.ipynb (19.12.2017)
  • 1D-DFT.ipynb (09.01.2018)
  • CWT.ipynb, cwt.py (10.01.2018)
  • DWT2.ipynb (10.01.2018)
  • YCrCbColorSpace.ipynb (16.01.2018)
  • PCA.ipynb (16.01.2018)
  • Eigenfaces.ipynb (17.01.2018)
  • LDA.ipynb (23.01.2018)
  • kNN.ipynb (24.01.2018)
  • Gaussians.ipynb (31.01.2018)
  • Just for fun: XKCD.ipynb, XKCDify.py

• If you are not able to work with IPython Notebook files, you can view the following html files. This is just a fallback solution since these html web sites are not interactive anymore: you cannot change the Python code and study the impact of the parameters, which is crucial for the understanding of the algorithms. However, you see the description, the program code, and the output.
  • GettingStarted.html
  • PlottingFunctions.html
  • FourierSeries.html
  • SawtoothDFT.html
  • Aliasing.html
  • WorkingWithImages.html
  • HistogramEqualization.html
  • ImageThresholding.html
  • Variance.html
  • UniformPDF.html
  • BoxFilterFrequencyResponse.html
  • FilteringImages.html
  • GradientVectorField.html
  • DetectingEdges.html
  • RecursiveFilter.html
  • MorphologicalOperations.html
  • CosineImages.html
  • WindowFunctions.html
  • 2D-DFT.html
  • 2D-DFT-Examples.html
  • WHT.html
  • 1D-DFT.html
  • CWT.html
  • DWT2.html
  • YCrCbColorSpace.html
  • PCA.html
  • Eigenfaces.html
  • LDA.html
  • kNN.html
  • Gaussians.html
  • Just for fun: XKCD.html