用几何引导的Mamba模块增强CNN分割，提升感受野且不增加太多计算量。
Generalizing Geometry-Guided Mamba as a Plug-and-Play Context Module for CNN-based Semantic Segmentation
- 将几何引导的Mamba嵌入现有CNN分割模型作为即插即用模块
- 在Cityscapes上实现稳定mIoU提升，额外计算量仅小幅增加
- 适合需要轻量化长程建模的视觉分割任务
基于CNN的语义分割网络通常依赖ASPP、PPM或注意力模块等上下文头来扩大感受野。这些模块虽有效，但可能带来高计算量、内存开销或边界泄漏问题。本文重新审视DGM-Net中的方向性几何Mamba（G-Mamba），将其作为即插即用的上下文聚合模块，而非完整新架构。核心思想是在选择性扫描过程中注入几何引导，使长距离特征传播受边界与向心流线索调控。我们替换了六种代表性CNN分割模型（DeepLabV3+、DANet、CCNet、PSPNet、PSANet、OCRNet）的原始上下文头，保持ResNet-101主干不变。在Cityscapes数据集上，$1024\times1024$分辨率下取得一致的mIoU提升，额外计算量仅适度增加，表明几何引导的SSM模块可作为传统上下文头的实际替代或增强方案。
原文摘要 · Abstract (English)
CNN-based semantic segmentation networks usually rely on context heads such as ASPP, PPM, or attention modules to enlarge the receptive field. These heads are effective but may introduce heavy computation, memory cost, or boundary leakage. This paper revisits Directional Geometric Mamba (G-Mamba) from DGM-Net and studies it as a plug-and-play context aggregation module rather than a complete new segmentation architecture. The key idea is to inject geometric guidance into the selective scan process, allowing long-range feature propagation to be modulated by boundary and centripetal-flow cues. We replace the original context heads of six representative CNN segmentation models, including DeepLabV3+, DANet, CCNet, PSPNet, PSANet, and OCRNet, while keeping the ResNet-101 backbone unchanged. Results on Cityscapes show consistent mIoU gains with only moderate extra GFLOPs at $1024\times1024$ resolution, suggesting that geometry-guided SSM modules can serve as practical alternatives or enhancements to conventional CNN context heads.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。