EM-SNN: Efficiently Modulated Spiking Neural Network for Remote Sensing Image Dehazing

2026-10-06

AI Summary

This study addresses the limitation of Spiking Neural Networks (SNNs) in remote sensing image dehazing, where the coupling between neuronal thresholds and haze degrades detail recovery. To overcome this, we propose EM-SNN, a novel framework that introduces the first statistics-driven TM-LIF neuron for adaptive threshold modulation and designs a Spike Sobel module to enhance structural feature representation. By leveraging event-driven sparse computation, the framework jointly optimizes activation scales and detail restoration. Experimental results demonstrate that EM-SNN achieves performance comparable to mainstream ANN-based baselines across multiple datasets while consuming only one-quarter of their energy. This work thereby enables high-accuracy, energy-efficient dehazing for remote sensing imagery.

Abstract

Although spiking neural networks (SNNs) provide an energy-efficient alternative to artificial neural networks (ANNs), their application to remote sensing image dehazing remains limited. A key challenge arises from the coupling between haze-induced high-frequency attenuation and discrete spike thresholding. This interaction suppresses weak responses and fundamentally limits the recovery of edges, textures, and fine details in spiking dehazing models. To address this challenge, we propose the Efficiently Modulated Spiking Neural Network (EM-SNN), a dedicated spiking framework tailored to remote sensing image dehazing. EM-SNN integrates a statistics-driven Threshold-Modulated Leaky Integrate-and-Fire (TM-LIF) neuron to adaptively compensate for haze-induced contrast compression, together with a Spike Sobel Modulation (SSM) module that enhances structural cues and reduces depth-wise attenuation during spiking feature propagation. By jointly modulating activation scales and structural representations, EM-SNN improves dehazing performance while preserving the inherent event-driven sparsity of SNNs. Experiments on HRSD, RICE, RRSHID, and SateHaze1K demonstrate that EM-SNN achieves competitive dehazing performance while consuming only one quarter of the energy of the strong ANN baseline SFRDP-Net.

Problem

• Spiking Neural Networks
• Remote Sensing Image Dehazing
• High-frequency Attenuation
• Spike Thresholding
• Detail Recovery

Innovation

• Spiking Neural Network
• Remote Sensing Image Dehazing
• Threshold-Modulated LIF Neuron
• Spike Sobel Modulation
• Energy Efficiency