Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.06920 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918377999564800 |
|---|---|
| author | Zhang, Qianqian Tabaro, Leon Abdelmoniem, Ahmed M. An, Junshe |
| author_facet | Zhang, Qianqian Tabaro, Leon Abdelmoniem, Ahmed M. An, Junshe |
| contents | Multispectral fusion object detection is a critical task for edge-based maritime surveillance and remote sensing, demanding both high inference efficiency and robust feature representation for high-resolution inputs. However, current State Space Models (SSMs) like Mamba suffer from significant parameter redundancy in their standard 2D Selective Scan (SS2D) blocks, which hinders deployment on resource-constrained hardware and leads to the loss of fine-grained structural information during conventional compression. To address these challenges, we propose the Low-Rank Two-Dimensional Selective Structured State Space Model (Low-Rank SS2D), which reformulates state transitions via matrix factorization to exploit intrinsic feature sparsity. Furthermore, we introduce a Structure-Aware Distillation strategy that aligns the internal latent state dynamics of the student with a full-rank teacher model to compensate for potential representation degradation. This approach substantially reduces computational complexity and memory footprint while preserving the high-fidelity spatial modeling required for object recognition. Extensive experiments on five benchmark datasets and real-world edge platforms, such as Raspberry Pi 5, demonstrate that our method achieves a superior efficiency-accuracy trade-off, significantly outperforming existing lightweight architectures in practical deployment scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_06920 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DLRMamba: Distilling Low-Rank Mamba for Edge Multispectral Fusion Object Detection Zhang, Qianqian Tabaro, Leon Abdelmoniem, Ahmed M. An, Junshe Computer Vision and Pattern Recognition Multispectral fusion object detection is a critical task for edge-based maritime surveillance and remote sensing, demanding both high inference efficiency and robust feature representation for high-resolution inputs. However, current State Space Models (SSMs) like Mamba suffer from significant parameter redundancy in their standard 2D Selective Scan (SS2D) blocks, which hinders deployment on resource-constrained hardware and leads to the loss of fine-grained structural information during conventional compression. To address these challenges, we propose the Low-Rank Two-Dimensional Selective Structured State Space Model (Low-Rank SS2D), which reformulates state transitions via matrix factorization to exploit intrinsic feature sparsity. Furthermore, we introduce a Structure-Aware Distillation strategy that aligns the internal latent state dynamics of the student with a full-rank teacher model to compensate for potential representation degradation. This approach substantially reduces computational complexity and memory footprint while preserving the high-fidelity spatial modeling required for object recognition. Extensive experiments on five benchmark datasets and real-world edge platforms, such as Raspberry Pi 5, demonstrate that our method achieves a superior efficiency-accuracy trade-off, significantly outperforming existing lightweight architectures in practical deployment scenarios. |
| title | DLRMamba: Distilling Low-Rank Mamba for Edge Multispectral Fusion Object Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.06920 |