港大开源新型语义分割框架SegMAN：Mamba和局部自注意力首次融进视觉主干
Ranking
Overall
47
Content
50
Popularity
41
Observed public metrics from 1 member.
Merged summary
TL;DR - HKU researchers introduced SegMAN, an open-source semantic segmentation framework combining Mamba-based global modeling with local self-attention and multiscale decoding. It improves segmentation accuracy and inference efficiency across major benchmarks while using fewer parameters than several larger vision backbones.
- The SegMAN Encoder’s LASS module combines sliding-window neighborhood attention for fine details with SS2D state-space scanning for long-range dependencies at linear complexity.
- Its Mamba-based MMSCopE decoder aggregates multiresolution features and propagates global semantics across scales.
- SegMAN-B’s 45M-parameter encoder reportedly outperforms larger VMamba-B and ConvNeXt-B backbones on ImageNet-1K.
- On 2048×1024 Cityscapes images, SegMAN-T achieves roughly three times EDAFormer-T’s frame rate while improving mIoU by 1.6 percentage points.
Sources (1)
港大开源新型语义分割框架SegMAN：Mamba和局部自注意力首次融进视觉主干
Public signals
Hugging Face upvotes 0
TL;DR - HKU researchers introduced SegMAN, an open-source semantic segmentation framework combining Mamba-based global modeling with local self-attention and multiscale decoding. It improves segmentation accuracy and inference efficiency across major benchmarks while using fewer parameters than several larger vision backbones.
- The SegMAN Encoder’s LASS module combines sliding-window neighborhood attention for fine details with SS2D state-space scanning for long-range dependencies at linear complexity.
- Its Mamba-based MMSCopE decoder aggregates multiresolution features and propagates global semantics across scales.
- SegMAN-B’s 45M-parameter encoder reportedly outperforms larger VMamba-B and ConvNeXt-B backbones on ImageNet-1K.
- On 2048×1024 Cityscapes images, SegMAN-T achieves roughly three times EDAFormer-T’s frame rate while improving mIoU by 1.6 percentage points.