How Far Can INRs Go? Cross-Domain Parameter-efficient INR-Based Semantic Segmentation for Brain MRI

2026-09-25

AI Summary

This study addresses the challenges of annotation scarcity, memory constraints, and cross-domain shift in brain MRI segmentation by investigating the performance boundaries of implicit neural representations (INRs) under low parameter budgets. Our analysis reveals that INRs exhibit significant advantages in few-shot scenarios and that semantic information is distributed across multiple layers. Accordingly, we propose HierINRSeg, a hierarchical architecture that enhances model robustness by aggregating multi-layer representations, while delineating its applicability conditions to guide model selection. Experimental results demonstrate that, compared to the MetaSeg baseline, our method improves Dice scores by 5.6% and 8.2% on in-domain and cross-domain evaluations, respectively. These findings validate the potential of INRs for lightweight and efficient medical image segmentation.

Abstract

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, we find that INR-based models do not simply improve with increasing parameter budget. Their advantage is most pronounced under low-parameter and limited-augmentation settings, while U-Net-based models benefit more from larger capacity and standard augmentation. We also investigate how INRs encode semantic information in their hidden features and show that complementary segmentation-relevant structure is distributed across multiple INR layers. Building on this insight, we introduce HierINRSeg, a hierarchical INR-based architecture that aggregates multi-layer representations for improved robustness and generalization. Extensive experiments show that HierINRSeg consistently outperforms MetaSeg, a strong recent INR-based segmentation baseline, with an average improvement of 5.6 percentage points in Dice for the in-domain test set and 8.2 percentage points out-of-domain. Overall, our analysis identifies the conditions under which INR-based segmentation is most effective, providing concrete guidance for model selection and future research.

Problem

• Implicit Neural Representations
• Semantic Segmentation
• Brain MRI
• Domain Generalization
• Parameter Efficiency

Innovation

• Implicit Neural Representations
• Semantic Segmentation
• Cross-Domain Generalization
• Hierarchical Architecture
• Parameter-Efficient