This code is used for generate sound, image, texture
The parameters there may not be optimal, so You may need to tune them. 
This code is only for research use, so cannot gurantee anything. Use with your own risk. 

release_object: call test_nfa.m to learning from object category
release_texture: call testTextureNFA.m to learn from texture category. (ususally just one image, but can also learn the fusion of two texture iages)
release_sound: call testAudioNFA.m to learn from sound signals. (Here, in the code, we try to mix two sound sources. its quite easy to adapt it to learn from just one sound source, but you need to modify the display a bit, otherwise it may give you the error. But the training process is quite the same)

