Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EventF2S: Asynchronous and Sparse Spiking AER Framework using Neuromorphic-Friendly Algorithm

Lakshmi Annamalai, Chetan Singh Thakur

TL;DR

EventF2S tackles the challenge of efficient, asynchronous processing of event-based AER data by introducing a Sparse Spiking Temporal Encoding (SS-TE) layer and a learnable First-To-Spike recognition network. The SS-TE layer reduces raw events to a single informative spike per pixel by leveraging parallel LIF filters and a contrast objective, enabling neuromorphic-friendly, low-power processing. The recognition component uses differentiable spiking dynamics with a first-spike protocol and a loss that encourages early correct-class firing, achieving competitive accuracy with significantly reduced computation. Together, the approach offers an asynchronous, sparse, and hardware-friendly solution for AER-based object recognition, well-suited for resource-constrained edge devices and neuromorphic hardware deployments.

Abstract

Bio-inspired Address Event Representation (AER) sensors have attracted significant popularity owing to their low power consumption, high sparsity, and high temporal resolution. Spiking Neural Network (SNN) has become the inherent choice for AER data processing. However, the integration of the AER-SNN paradigm has not adequately explored asynchronous processing, neuromorphic compatibility, and sparse spiking, which are the key requirements of resource-constrained applications. To address this gap, we introduce a brain-inspired AER-SNN object recognition solution, which includes a data encoder integrated with a First-To-Spike recognition network. Being fascinated by the functionality of neurons in the visual cortex, we designed the solution to be asynchronous and compatible with neuromorphic hardware. Furthermore, we have adapted the principle of denoising and First-To-Spike coding to achieve optimal spike signaling, significantly reducing computation costs. Experimental evaluation has demonstrated that the proposed method incurs significantly less computation cost to achieve state-of-the-art competitive accuracy. Overall, the proposed solution offers an asynchronous and cost-effective AER recognition system that harnesses the full potential of AER sensors.

EventF2S: Asynchronous and Sparse Spiking AER Framework using Neuromorphic-Friendly Algorithm

TL;DR

EventF2S tackles the challenge of efficient, asynchronous processing of event-based AER data by introducing a Sparse Spiking Temporal Encoding (SS-TE) layer and a learnable First-To-Spike recognition network. The SS-TE layer reduces raw events to a single informative spike per pixel by leveraging parallel LIF filters and a contrast objective, enabling neuromorphic-friendly, low-power processing. The recognition component uses differentiable spiking dynamics with a first-spike protocol and a loss that encourages early correct-class firing, achieving competitive accuracy with significantly reduced computation. Together, the approach offers an asynchronous, sparse, and hardware-friendly solution for AER-based object recognition, well-suited for resource-constrained edge devices and neuromorphic hardware deployments.

Abstract

Bio-inspired Address Event Representation (AER) sensors have attracted significant popularity owing to their low power consumption, high sparsity, and high temporal resolution. Spiking Neural Network (SNN) has become the inherent choice for AER data processing. However, the integration of the AER-SNN paradigm has not adequately explored asynchronous processing, neuromorphic compatibility, and sparse spiking, which are the key requirements of resource-constrained applications. To address this gap, we introduce a brain-inspired AER-SNN object recognition solution, which includes a data encoder integrated with a First-To-Spike recognition network. Being fascinated by the functionality of neurons in the visual cortex, we designed the solution to be asynchronous and compatible with neuromorphic hardware. Furthermore, we have adapted the principle of denoising and First-To-Spike coding to achieve optimal spike signaling, significantly reducing computation costs. Experimental evaluation has demonstrated that the proposed method incurs significantly less computation cost to achieve state-of-the-art competitive accuracy. Overall, the proposed solution offers an asynchronous and cost-effective AER recognition system that harnesses the full potential of AER sensors.
Paper Structure (17 sections, 6 equations, 1 figure, 3 tables)

This paper contains 17 sections, 6 equations, 1 figure, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Sparse Temporal-Encoded EventF2S Architecture
  4. Single Spike Temporal Encoding (SS-TE) Layer
  5. SS-TE Encoding Function
  6. Learning by First-To-Spike Spiking Network
  7. Neuron Dynamics
  8. Loss Function
  9. Experiments and Results
  10. SS-TE Experiments
  11. Qualitative Analysis
  12. Denoising Experiments
  13. Sparse Temporal Spiking
  14. Recognition Experiments
  15. comparison in terms of Recognition Accuracy
  16. ...and 2 more sections

Figures (1)

  • Figure 1: Left: Raw event data of digit $5$ drawn from neuromorphic NMNIST dataset. Right: SS-TE encoding. SS-TE output being visually recognizable reinstates that SS-TE retains necessary information with just a single spike per pixel.