Hopfield neuro network
Description
This project is naive (and, therefore, simple!) implementation of Hopfield neural network. The goals for this project are:
- Keep implementation simple, easy to understant and free for further experiments
- Show basics how neural network work and learns
- Follow strong math basics described by G.G. Matinetsky (see links)
How to run
There are 3 'reference' images and 2 'noisy' images. See 'images_to_learn' and 'noisy_examples' folders for details.
Example:
$ ./hopfield 10 10 ../noisy_example/src.txt ../images_to_learn/*.txt
How to build
This project is based on CMAKE build system (cmake version 3+). So, just execute:
Ubuntu
$ mkdir build ; cd build
$ cmake ..
$ make
Windows (Visual Studio)
- Download and install CMAKE
- Open console at repository folder
- $ mkdir build ; cd build
- $ cmake ..
- Open generated solution with Visual Studio and compile it ordinal way
Links:
- Г.Г. Малинецкий. Математические основы синергетики. Москва, URSS, 2009 [In Russian]