The goal of this work is to replicate the neural systems that process visual images in the human eyes and brain. The Images enter the software through a webcam, and are then processed using spiking convolutional neural networks that simulate Gabor filtering and Max pooling operations. In higher layers STDP learning is used. The software is written in Java and can simulate millions of neurons on a laptop.
The image below shows the neuron voltages in color. The horizontal axis is the neuron number and the vertical axis is time. The vertical yellow stripe represents a neuron that is not firing.
Image credit: Lyle N. Long, APS Fellow, Pennsylvania State University, University Park, PA. Research was done while Dr. Long was on sabbatical at California Institute of Technology, Pasadena, CA