SpikeRestormer: Towards Energy-Efficient All-in-One Image Restoration via Unified Event Reasoning

2026-08-03

AI Summary

This work addresses the challenge of efficient universal image restoration by introducing SpikeRestormer, the first spike neural network (SNN)-based framework tailored for this task. While conventional approaches rely on artificial neural networks (ANNs) with high computational costs that hinder real-time deployment, SNNs struggle with static images due to the absence of explicit event signals and the tight coupling between degradation and image structure. SpikeRestormer overcomes these limitations by unifying event-based reasoning for static inputs, formulating restoration as a joint process of degradation event perception, reliability inference, and restoration event construction. Its core components—Subtractive Degradation Event Attention (SDEA), Hierarchical Bayesian Skip Masking (HBSM), and Additive Restoration Event Attention (AREA)—enable it to match ANN-level restoration performance while achieving state-of-the-art results among SNNs and significantly reducing energy consumption, thus striking an excellent balance between performance and efficiency.

Abstract

ANN-based All-in-One image restoration (AiOIR) unifies diverse degradation handling but incurs high computational costs, limiting its real-time deployment. While Spiking Neural Networks (SNNs) offer a low-power alternative, applying them to static images remains challenging. This difficulty arises because explicit event signals are absent, and degradation cues are heavily entangled with scene structures, hindering the learning of reliable restoration-oriented spike events. To address these issues, we propose SpikeRestormer, an energy-efficient SNN for AiOIR that performs event reasoning over internally generated spike cues. Specifically, we propose a degradation-event perception process to extract spike-based degradation events through Subtractive Degradation Event Attention (SDEA). Moreover, we introduce Hierarchical Bayesian Skip Masking (HBSM) and Additive Restoration Event Attention (AREA) processes for event-reliability inference and restoration-event construction, respectively. By integrating these complementary processes, SpikeRestormer formulates restoration as a unified process of degradation-event perception, degradation-event reliability inference, and restoration-event construction, liberating the potential of SNNs for energy-efficient AiOIR. Extensive experiments show that SpikeRestormer delivers competitive performance against ANN-based methods and establishes new state-of-the-art results among SNN-based methods with significantly lower energy consumption.

Problem

• image restoration
• spiking neural networks
• energy efficiency
• event reasoning
• degradation cues

Innovation

• Spiking Neural Networks
• All-in-One Image Restoration
• Event Reasoning
• Degradation-Event Perception
• Energy Efficiency