/advanced-image-processing-and-computer-vision-ps

Primary LanguageOpenEdge ABLMIT LicenseMIT

IRS - Image Recognition System

Developed as main part of the proseminar (pratical part) of the "Advanced Image Processing & Computer Vision" class at the Department of Computer Sciences, University of Salzburg, Austria.

Requirements

The code was tested using the following setup:

  • 4x3.60 GHz AMD FX-4100 CPU, 12GB main memory, Linux Mint 17.
  • python in version 2.7.6

Some directories have to be present, see report and src/settings/settings.py.

Remarks

The k-means++ clustering uses 4 cores in parallel (by argument), so better make sure to change this parameter (in src/kmeanspp/kmeanspp.py) or have 4 cores available (I did not investigate how the scikit-learn kMeans behaves if there are less than 4 cores available).