Spike-Based Approximate Backpropagation Algorithm of Brain-Inspired Deep SNN for Sonar Target Classification

<div>Overall architecture of brain-inspired SNNs. The encoder converts the input images into spikes, the hidden layers consist of LIF neurons, which can be any network structure (such as LeNet, VGG, and ResNet), and the decoder converts the output spikes into the corresponding classification results.</div>

Computational Intelligence and Neuroscience

fig1

Figure 1

Figure 1: Spike-Based Approximate Backpropagation Algorithm of Brain-Inspired Deep SNN for Sonar Target Classification 