FSB-Net: Frequency-Spatial Boundary Network for Brain Stroke Lesion Segmentation in Non-Contrast CT

2026-07-23

AI Summary

This study addresses the significant challenge of segmenting stroke lesions in non-contrast CT scans, where low contrast, morphological heterogeneity, and ambiguous boundaries hinder accurate delineation, and existing methods lack explicit boundary modeling. To overcome these limitations, this work proposes FSB-Net, the first framework to integrate wavelet transform and spectral boundary loss for stroke segmentation. FSB-Net leverages discrete wavelet transform to extract multi-scale boundary features and introduces a Wavelet Boundary Detection Head along with a Frequency-Spatial Cross-Attention Module to effectively fuse frequency- and spatial-domain information. Furthermore, a Spectral Boundary Loss is designed to enhance boundary precision. Evaluated on public datasets, FSB-Net consistently outperforms state-of-the-art models—including U-Net, UNet++, MANet, and DeepLabV3+—achieving leading performance across Dice, IoU, and HD95 metrics.

Abstract

Accurate segmentation of brain stroke lesions in non-contrast computed tomography (NCCT) scans is critical for rapid clinical decision-making, yet remains difficult due to the low contrast between lesion and normal brain tissue, heterogeneous lesion morphology across ischemic and hemorrhagic subtypes, and ambiguous boundaries caused by partial volume effects. Current deep learning approaches primarily optimize region-level overlap but lack explicit boundary modeling, leading to imprecise delineation that can affect volumetric assessment and treatment planning. We propose FSB-Net, a frequency-spatial boundary network that leverages frequency-domain analysis for boundary-aware stroke lesion segmentation. FSB-Net introduces three components: (i) a Wavelet Boundary Detection Head (WBDH) that applies the discrete wavelet transform to multi-scale encoder features, extracting high-frequency sub-bands as boundary representations; (ii) a Frequency-Spatial Cross-Attention Module (FSCAM) that performs bidirectional attention between wavelet boundary features and spatial decoder features for selective boundary enhancement; and (iii) a Spectral Boundary Loss that penalizes high-frequency discrepancies in the Fourier domain to optimize boundary sharpness. Built on a PVTv2-B2 encoder, FSB-Net is evaluated on a public Brain Stroke CT dataset containing both ischemic and hemorrhagic cases. Experimental results show that FSB-Net outperforms U-Net, UNet++, MANet, and DeepLabV3+ across all metrics, achieving state-of-the-art performance in mean Dice, mean IoU, and HD95.

Problem

• stroke lesion segmentation
• non-contrast CT
• boundary ambiguity
• low contrast
• heterogeneous morphology

Innovation

• frequency-spatial boundary
• wavelet transform
• cross-attention
• boundary-aware segmentation
• non-contrast CT