[论文] How Far Can INRs Go? Cross-Domain Parameter-efficient INR-Based Semant...

论文概要

研究领域: CV 作者: Ziyao Shang, Pouya Sadeghi, Letian Jiang, Alexander Wong, Sirisha Rambhatla 发布时间: 2026-09-25 arXiv: 2609.31573

中文摘要

生物医学图像分割是医学图像分析的核心，但实际部署常面临标注有限、内存受限和跨站点分布漂移等挑战。隐式神经表示（INR）近来作为一种轻量级的语义分割替代方案出现，以显著更少的参数取得有竞争力的性能。然而，INR 分割的机制、扩展行为与域泛化能力仍缺乏充分理解。本文在跨域脑 MRI 分割的语境下研究这些问题。我们在低参数区间分析基于 INR 的分割，并与传统流程在域内和域外两种设定下比较。令人意外的是，我们发现 INR 模型并非随参数预算增加而单调改善：其优势在低参数与有限增广设定下最为明显，而基于 U-Net 的模型则更受益于更大容量和标准增广。我们还研究了 INR 如何在其隐藏特征中编码语义信息，证明互补的分割相关结构分布在多个 INR 层中。基于这一洞见，我们提出 HierINRSeg——一种分层的 INR 分割架构，聚合多层表示以提升鲁棒性与泛化能力。大量实验表明，HierINRSeg 持续优于近期强基线 MetaSeg：域内测试集 Dice 平均提升 5.6 个百分点，域外提升 8.2 个百分点。总体而言，我们的分析识别了 INR 分割最有效的条件，为模型选择与未来研究提供了具体指导。

原文摘要

Biomedical image segmentation is central to medical image analysis, but practical deployment often faces limited annotations, memory constraints, and cross-site distribution shifts. Implicit Neural Representations (INRs) have recently emerged as a lightweight alternative for semantic segmentation, achieving competitive performance with substantially fewer parameters than conventional architectures. However, the mechanisms, scaling behavior, and domain generalization abilities of INR-based segmentation remain insufficiently understood. In this work, we study these questions in the context of cross-domain brain MRI segmentation. We analyze INR-based segmentation across low-parameter regimes, comparing it with conventional pipelines in both in-domain and out-of-domain settings. Surprisingly, ...

#论文 #arXiv #CV #小凯