STDP-Based Spiking Deep Convolutional Neural Networks for Object Recognition

By Saeed Reza Kheradpisheh et al

Published on June 10, 2017

Read the original document by opening this link in a new tab.

Table of Contents

Abstract
Introduction
Proposed Spiking Deep Neural Network
DoG and Temporal Coding
Convolutional Layers
Local Pooling Layers
STDP-based Learning

Summary

This paper presents a study on STDP-based spiking deep convolutional neural networks for object recognition. The authors propose a deep SNN architecture that utilizes spike-timing-dependent plasticity (STDP) for unsupervised learning of visual features. The network comprises several convolutional and pooling layers, designed to progressively learn and detect object prototypes. Through experiments on various datasets, the proposed network demonstrates high accuracy and energy efficiency. The paper highlights the advantages of STDP over other unsupervised techniques and discusses the potential implications for artificial vision systems.