Python Machine Learning OpenCV Teaching Examples

OpenCV Python machine learning examples used for teaching within the undergraduate Computer Science programme at Durham University (UK) by Prof. Toby Breckon.

All tested with OpenCV 3.x and Python 3.x.

---

Background:

Directly adapted from the older C++ OpenCV machine learning teaching examples (that for a long time, due to the absence of other fully worked examples for the OpenCV machine learning components became the de facto reference for the use of these OpenCV routines)

All dataset examples are taken and reproduced from the UCI Machine Learning Repository.

A related set of Python Image Processing OpenCV Teaching Examples are also available covering basic image processing operations and similarly a set of Python Computer Vision OpenCV Teaching Examples

---

How to download and run:

Download each file as needed or to download the entire repository and run each try:

git clone https://github.com/tobybreckon/python-examples-ml.git
cd python-examples-ml
cd <sub directory of one of the examples>
python3 ./<insert file name of one of the examples>.py

-- which should then produce an output of the results on the dataset inside that example directory.

In each sub-directory:

• .py file(s) - code for the examples (several examples per directory in many cases)
• .name file - an explanation of the data and its source
• .data file - the original and complete set of data (CSV file format)
• .train file - the data to be used for training (CSV file format)
• .test file - the data to be used for testing (CSV file format)
• .xml, .yml - if present, example data files for testing some tools

For some examples you may need to copy/link the .train/.test files from one of the other directories (it seemed silly to archive them multiple times).

Demo source code is provided "as is" to aid learning and understanding of topics on the course and beyond.

---

If referencing these examples in your own work please use:

@TechReport{breckon2014,
  author =       {Breckon, T.P.},
  title =        {Machine Learning - Course Notes and Materials},
  institution =  {Durham University},
  year =         {2014},
  address =      {Durham, UK},
}

---

If you find any bugs raise an issue (or much better still submit a git pull request with a fix) - toby.breckon@durham.ac.uk

"may the source be with you" - anon.