The author introduces a Recurrent Spiking Transformer framework to efficiently extract spatio-temporal features for visual saliency detection while minimizing power consumption.
