MADSNet：多尺度轴向注意力密集脉冲网络用于脉冲相机图像重建
MADSNet: A Multi-Scale Axial-Attentive Dense Spiking Network for Spike Camera Image Reconstruction
AI 导读
MADSNet 是一种用于脉冲相机图像重建的多尺度轴向注意力密集脉冲网络（SNN），采用分层脉冲编码器-解码器结构，包含多尺度脉冲 stem、轴向脉冲注意力、密集脉冲瓶颈和 spike-to-continuous 解码器。在 SREDS 上，MADSNet 比 SSIR 高 4.3 dB PSNR，比 Spk2ImgNet 高 1.94 dB，且延迟约低 15 倍、能耗仅约 1/40。
正文 · 原文
Abstract:Spike cameras encode dynamic scenes into high-temporal-resolution binary spike streams for efficient high-speed vision. Although spiking neural networks (SNNs) can process such signals with sparse and accumulate-only computation, existing SNN-based reconstruction methods are typically shallow and still lag behind ANN-based models in reconstruction fidelity. This paper proposes MADSNet, a Multi-scale Axial-attentive Dense Spiking Network for spike camera image reconstruction. MADSNet adopts a hierarchical spiking encoder--decoder, where a multi-scale spike stem extracts early features, axial spike attention captures long-range dependencies, a dense spiking bottleneck enhances feature reuse, and a spike-to-continuous decoder reconstructs continuous intensity images from spike-domain representations. Experiments on gray-scale and color spike reconstruction show that MADSNet outperforms SSIR by 4.3 dB PSNR and Spk2ImgNet by 1.94 dB on SREDS. Compared with Spk2ImgNet, it achieves about 15$\times$ lower latency and uses only about 1/40 of the estimated energy, demonstrating that carefully designed SNN architectures can improve reconstruction quality while preserving computational efficiency.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.12980 [cs.LG] |
| (or arXiv:2611.12980v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.12980 arXiv-issued DOI via DataCite |
Submission history
From: Sirui Li [view email]
[v1]
Sun, 8 Nov 2026 07:05:27 UTC (7,950 KB)
来源：arXiv cs.AI · arxiv.org