Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bilateral Hippocampi Segmentation in Low Field MRIs Using Mutual Feature Learning via Dual-Views

Himashi Peiris, Zhaolin Chen

TL;DR

A novel deep-learning approach for the automatic segmentation of bilateral hippocampi in low-field MRIs, which employs a dual-view structure to enable mutual feature learning via high-frequency masking, enhancing segmentation accuracy by leveraging complementary information from different perspectives.

Abstract

Accurate hippocampus segmentation in brain MRI is critical for studying cognitive and memory functions and diagnosing neurodevelopmental disorders. While high-field MRIs provide detailed imaging, low-field MRIs are more accessible and cost-effective, which eliminates the need for sedation in children, though they often suffer from lower image quality. In this paper, we present a novel deep-learning approach for the automatic segmentation of bilateral hippocampi in low-field MRIs. Extending recent advancements in infant brain segmentation to underserved communities through the use of low-field MRIs ensures broader access to essential diagnostic tools, thereby supporting better healthcare outcomes for all children. Inspired by our previous work, Co-BioNet, the proposed model employs a dual-view structure to enable mutual feature learning via high-frequency masking, enhancing segmentation accuracy by leveraging complementary information from different perspectives. Extensive experiments demonstrate that our method provides reliable segmentation outcomes for hippocampal analysis in low-resource settings. The code is publicly available at: https://github.com/himashi92/LoFiHippSeg.

Bilateral Hippocampi Segmentation in Low Field MRIs Using Mutual Feature Learning via Dual-Views

TL;DR

A novel deep-learning approach for the automatic segmentation of bilateral hippocampi in low-field MRIs, which employs a dual-view structure to enable mutual feature learning via high-frequency masking, enhancing segmentation accuracy by leveraging complementary information from different perspectives.

Abstract

Accurate hippocampus segmentation in brain MRI is critical for studying cognitive and memory functions and diagnosing neurodevelopmental disorders. While high-field MRIs provide detailed imaging, low-field MRIs are more accessible and cost-effective, which eliminates the need for sedation in children, though they often suffer from lower image quality. In this paper, we present a novel deep-learning approach for the automatic segmentation of bilateral hippocampi in low-field MRIs. Extending recent advancements in infant brain segmentation to underserved communities through the use of low-field MRIs ensures broader access to essential diagnostic tools, thereby supporting better healthcare outcomes for all children. Inspired by our previous work, Co-BioNet, the proposed model employs a dual-view structure to enable mutual feature learning via high-frequency masking, enhancing segmentation accuracy by leveraging complementary information from different perspectives. Extensive experiments demonstrate that our method provides reliable segmentation outcomes for hippocampal analysis in low-resource settings. The code is publicly available at: https://github.com/himashi92/LoFiHippSeg.

Paper Structure

This paper contains 22 sections, 11 equations, 3 figures, 4 tables.

Table of Contents

  1. Introdcution
  2. Dataset
  3. Methodology
  4. Notations & Problem Formulation.
  5. LoFiHippSeg Architecture.
  6. Frequency Masking Module (FMM).
  7. Objective Function.
  8. The Segmentation Loss ($\mathop{\mathrm{\mathcal{L}}}\nolimits_{\mathrm{SEG}}$).
  9. The Adversarial Loss ($\mathop{\mathrm{\mathcal{L}}}\nolimits_{\mathrm{ADV}}$).
  10. The Masked Spatial Loss ($\mathop{\mathrm{\mathcal{L}}}\nolimits_{\mathrm{M}}$).
  11. The Critic Loss ($\mathop{\mathrm{\mathcal{L}}}\nolimits_{\mathrm{C}}$).
  12. Experiments
  13. Implementation Details
  14. Evaluation Metrics
  15. Experimental Results
  16. ...and 7 more sections

Figures (3)

  • Figure 1: Sample case from LISA Dataset.
  • Figure 2: Overview of Proposed Dual-View Pipeline, LoFiHippSeg. Here, $\mathcal{F}_1(\cdot)$ and $\mathcal{F}_2(\cdot)$ are structuraly similar VNet models.
  • Figure 3: Feature difference between Low-field Image and High-frequency Image.