Adaptive resonance theory (ART) describes several neural network models for pattern recognition and prediction. This project is using ART1 neural network type for clustering black & white images as described in the attached documentation.
The neural network is sensitive to "vigilance" parameter. Higher vigilance produces highly detailed clusters, while lower vigilance results in more general clusters.
The program is expecting black & white PNG pictures in RGBA representation, where only the alpha (4th bit) determines whether the image is black or white.
Use ant to compile and run the program's GUI.
ant clean
ant compile (includes clean)
ant jar (includes compile)
ant run (includes jar)
ant (same as ant run)
The project was created as a homework within Soft Computing course at Faculty of Information Technology, Brno University of Technology, 2014.