scroll down for instructions and explanations

This applet implements a solution of the image segmentation problem using the LEGION artificial neural network.

Click on the grid to form a black and white image consisting of no more than 3 disconnected parts (such as the word "C U P"). If you want to trace the activity of a particular segment, you clicking twice will create a "tracer" cell (colored red). It is suggested that you have one tracer per individual segment (e.g. each letter in the word C. U. P.), since cells within the same segment behave similarly. After you are done creating the image and setting up tracers, press the Run button, and you will see the network's oscillators become active. Eventually neurons within the same segment synchronize with each other while desynchronizing from neurons that belong to other disconnected segments. The final result is not a stable configuration, but rather a set of waves of activity alternating between disconnected regions. Thus the image is segmented temporally, the behavior of each region alternating between active (excited) and inactive (inhibited) states.

For more information refer to: The network design is based on the paper by Wang D.L. (1995): Emergent synchrony in locally coupled neural oscillators. IEEE Transactions on Neural Networks, vol. 6, 941-948.