用自适应三视图融合让轻量CNN分割更准
ATV-Net: Adaptive Triple-View Network with Dynamic Feature Fusion
- 三视图感知：点、邻域、上下文，动态加权融合
- Cityscapes上达80.31% mIoU，GFLOPs更低
- 适合追求高效精准的视觉任务开发者
近期语义分割依赖基于注意力和Transformer的架构，虽精度高但结构复杂、计算开销大。本文探讨轻量级CNN分割头能否通过自适应选择有效感受野信息保持竞争力。提出ATV-Net，一种在常规骨干网络上附加的轻量级分割头，整合点级、邻域级与扩大上下文三种互补视图，并通过自适应决策门生成依赖图像的权重，根据场景内容动态强调不同感受野响应，无需密集注意力或多尺度聚合。在Cityscapes和Pascal VOC 2012上的实验表明，使用ResNet-101时，ATV-Net在Cityscapes上达到80.31% mIoU，使用ConvNeXt-Tiny时达80.90%；在Pascal VOC 2012上分别达到86.7%和88.5% mIoU，且所需GFLOPs低于代表性上下文聚合与注意力头。结果表明，自适应感受野选择是CNN语义分割中实用有效的设计路径。
原文摘要 · Abstract (English)
Recent advances in semantic segmentation rely heavily on attention-based and transformer-style architectures that, while accurate, introduce considerable architectural complexity and computational cost. This paper asks whether a compact CNN-based segmentation head can remain competitive by adaptively selecting useful receptive-field evidence. We propose ATV-Net, an Adaptive Triple-View Network that attaches a lightweight head to a conventional backbone. The head organizes three complementary views -- point-wise, neighborhood-level, and enlarged context -- and fuses them through an Adaptive Decision Gate that generates image-dependent weights from global feature statistics. This allows the model to emphasize different receptive-field responses according to scene content, without dense attention or multi-scale aggregation. Experiments on Cityscapes and Pascal VOC 2012 show that ATV-Net achieves 80.31% mIoU on Cityscapes with ResNet-101 and 80.90% with ConvNeXt-Tiny, and 86.7% and 88.5% mIoU on Pascal VOC 2012, respectively, while requiring fewer GFLOPs than representative context-aggregation and attention-based heads. The results indicate that adaptive receptive-field selection remains a practical and effective design choice for CNN-based semantic segmentation.
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。