Ultralight Polarity-Split Neuromorphic SNN for Event-Stream Super-Resolution

TL;DR

This paper introduces an ultra-lightweight, event-stream super-resolution method using Spiking Neural Networks, featuring a novel polarity-split encoding and a learnable loss, enabling real-time, resource-efficient perception enhancement.

Contribution

The work presents a novel polarity-split encoding strategy and a learnable loss function for SNN-based event super-resolution, optimizing for lightweight and real-time deployment.

Findings

Achieves competitive super-resolution performance on multiple datasets.

Significantly reduces model size and inference time.

Enables embedding into event cameras for real-time applications.

Abstract

Event cameras offer unparalleled advantages such as high temporal resolution, low latency, and high dynamic range. However, their limited spatial resolution poses challenges for fine-grained perception tasks. In this work, we propose an ultra-lightweight, stream-based event-to-event super-resolution method based on Spiking Neural Networks (SNNs), designed for real-time deployment on resource-constrained devices. To further reduce model size, we introduce a novel Dual-Forward Polarity-Split Event Encoding strategy that decouples positive and negative events into separate forward paths through a shared SNN. Furthermore, we propose a Learnable Spatio-temporal Polarity-aware Loss (LearnSTPLoss) that adaptively balances temporal, spatial, and polarity consistency using learnable uncertainty-based weights. Experimental results demonstrate that our method achieves competitive super-resolution performance on multiple datasets while significantly reducing model size and inference time. The lightweight design enables embedding the module into event cameras or using it as an efficient front-end preprocessing for downstream vision tasks.