通过注意力机制融合局部细节与全局语义，提升无监督域适应分割精度
AFRDA: Attentive Feature Refinement for Domain Adaptive Semantic Segmentation
- 利用低分辨率预测生成语义先验，精炼高分辨率特征
- 引入高频成分捕捉边缘信息，改善物体边界分割
- 不确定性驱动注意力自适应平衡局部与全局信息，适合复杂场景
在无监督域自适应语义分割（UDA-SS）中，模型在带标注的源域数据（如合成图像）上训练，并适应到无标注的目标域（如真实世界图像）。现有方法常难以兼顾细粒度局部细节与全局上下文信息，导致复杂区域出现分割错误。为此，本文提出自适应特征精炼（AFR）模块，通过低分辨率预测的语义先验来优化高分辨率特征，提升分割精度；同时引入高频分量以捕获细粒度结构并提供关键边界信息，增强物体轮廓刻画；此外，基于不确定性驱动的注意力机制自适应平衡局部与全局信息，减少误分类。其轻量化设计可无缝集成至基于HRDA的UDA方法中，实现当前最优性能：在GTA V → Cityscapes任务上提升1.05% mIoU，Synthia → Cityscapes任务上提升1.04% mIoU。代码已开源。
原文摘要 · Abstract (English)
In Unsupervised Domain Adaptive Semantic Segmentation (UDA-SS), a model is trained on labeled source domain data (e.g., synthetic images) and adapted to an unlabeled target domain (e.g., real-world images) without access to target annotations. Existing UDA-SS methods often struggle to balance fine-grained local details with global contextual information, leading to segmentation errors in complex regions. To address this, we introduce the Adaptive Feature Refinement (AFR) module, which enhances segmentation accuracy by refining highresolution features using semantic priors from low-resolution logits. AFR also integrates high-frequency components, which capture fine-grained structures and provide crucial boundary information, improving object delineation. Additionally, AFR adaptively balances local and global information through uncertaintydriven attention, reducing misclassifications. Its lightweight design allows seamless integration into HRDA-based UDA methods, leading to state-of-the-art segmentation performance. Our approach improves existing UDA-SS methods by 1.05% mIoU on GTA V --> Cityscapes and 1.04% mIoU on Synthia-->Cityscapes. The implementation of our framework is available at: https://github.com/Masrur02/AFRDA
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。