Research map / Detection and segmentation / CNN features
MambaSeg: Harnessing Mamba for Accurate and Efficient Image-Event Semantic Segmentation
Fuqiang Gu, Yuanke Li, Xianlei Long, Kangping Ji, Chao Chen, Qingyi Gu, Zhenliang Ni
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use DDIM fusion instead of element-wise addition (ablation) | MambaSeg (VMamba-T) · DDD17 | mIoU (%): 74.38 -> 77.56 | Improved |
| use DDIM instead of FFM (ablation) | MambaSeg · DDD17 | mIoU (%): 76.44 -> 77.56 | Improved |
| use DDIM instead of CSF (ablation) | MambaSeg · DDD17 | mIoU (%): 76.65 -> 77.56 | Improved |
| add CSIM (ablation) | MambaSeg with CTIM · DDD17 | mIoU (%): 76.20 -> 77.56 | Improved |
| add CTIM (ablation) | MambaSeg with CSIM · DDD17 | mIoU (%): 76.32 -> 77.56 | Improved |
| add SS2D to CSIM (ablation) | CSIM · DDD17 | mIoU (%): 76.71 -> 77.56 | Improved |
| add cross spatial attention to CSIM (ablation) | CSIM · DDD17 | mIoU (%): 76.59 -> 77.56 | Improved |
| add spatial attention to CSIM (ablation) | CSIM · DDD17 | mIoU (%): 76.47 -> 77.56 | Improved |
| add BTSS to CTIM (ablation) | CTIM · DDD17 | mIoU (%): 76.53 -> 77.56 | Improved |
| add temporal attention to CTIM (ablation) | CTIM · DDD17 | mIoU (%): 76.49 -> 77.56 | Improved |
| add cross temporal attention to CTIM (ablation) | CTIM · DDD17 | mIoU (%): 76.66 -> 77.56 | Improved |
| Mamba backbone with DDIM instead of Transformer fusion (EISNet) | MambaSeg · DDD17 | Params (M): 34.39 -> 25.44 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SegFormer — SegFormer: A Topic Segmentation Model with Controllable Range of Attention (AAAI 2023) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Efficient Event-Based Semantic Segmentation via Exploiting Frame-Event Fusion: A Hybrid Neural Network Approach (AAAI 2025 · similarity 0.88)
- Mamba-Reg: Vision Mamba Also Needs Registers (CVPR 2025 · similarity 0.87)
- Mamba as a Bridge: Where Vision Foundation Models Meet Vision Language Models for Domain-Generalized Semantic Segmentation (CVPR 2025 · similarity 0.86)
- EventMamba: Enhancing Spatio-Temporal Locality with State Space Models for Event-Based Video Reconstruction (AAAI 2025 · similarity 0.86)
- ESEG: Event-Based Segmentation Boosted by Explicit Edge-Semantic Guidance (AAAI 2025 · similarity 0.86)
- Re-coding for Uncertainties: Edge-awareness Semantic Concordance for Resilient Event-RGB Segmentation (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
