Detection of Straight Lines Using a Spiking Neural Network Model

Receptive fields of neurons play various roles in biological neural networks. Based on a receptive field with the function of Hough transform, a spiking neural network model is proposed to detect straight lines in a visual image. Through the network, straight lines transform to corresponding neurons with high firing rates in the output neuron array. Simulation results show that straight lines can be detected by the network and firing rates of the corresponding neurons are referred to lengths of the lines. This model can be used to explain how a spiking neuron-based network can detect straight lines, and furthermore the model can be used in an artificial intelligent system.