Neocognitron Simulator Screenshots

Fig. 1: Neocognitron simulator's main window

Fig. 2: Window to enter (draw, load or save) input layer U0

Fig. 3: In this dialog, you can set the input panel (U0) size accoring to the number of layers

Fig. 4: You can manage (draw new/load/save/skip) patterns to training

Fig. 5: When you prepared patterns, and adjusted the Neocognitron's structure, you are ready to learn the NeoCognitron
Remark: The size of patterns must equals to the input panel size (U0)

Fig. 6: The Result window informs you about the winner output, and it shows the corresponding trained pattern.

Fig. 7: Right-click on panel to view the corresponding wieghts

Fig. 8: Cell values on panel: Us3[3]

Fig. 9: Save weights - context menu Note: The appliacation is able to save the trained neural nets, with patterns used to learn too.

Fig. 10: Trained A weights between Uc2[3] and Us3[1]