Event-based Optical Flow on Neuromorphic Processor: ANN vs. SNN Comparison based on Activation Sparsification

TL;DR

This paper compares the efficiency of ANN and SNN for event-based optical flow on a neuromorphic processor, demonstrating SNN's superior energy and time efficiency due to lower spike density and activation sparsity.

Contribution

It introduces a sparsification-aware training method and a hardware-in-loop evaluation to fairly compare ANN and SNN performance on neuromorphic hardware.

Findings

SNN consumes 62.5% of the time and 75.2% of the energy of ANN.

Both networks achieve similar low activation/spike density (~5%).

SNN's higher efficiency is due to lower spike density reducing memory access.

Abstract

Spiking neural networks (SNNs) for event-based optical flow are claimed to be computationally more efficient than their artificial neural networks (ANNs) counterparts, but a fair comparison is missing in the literature. In this work, we propose an event-based optical flow solution based on activation sparsification and a neuromorphic processor, SENECA. SENECA has an event-driven processing mechanism that can exploit the sparsity in ANN activations and SNN spikes to accelerate the inference of both types of neural networks. The ANN and the SNN for comparison have similar low activation/spike density (~5%) thanks to our novel sparsification-aware training. In the hardware-in-loop experiments designed to deduce the average time and energy consumption, the SNN consumes 44.9ms and 927.0 microjoules, which are 62.5% and 75.2% of the ANN's consumption, respectively. We find that SNN's higher efficiency attributes to its lower pixel-wise spike density (43.5% vs. 66.5%) that requires fewer memory access operations for neuron states.