A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques

Hippocampus segmentation on MRI (magnetic resonance imaging) plays a vital role in detecting, diagnosing, tracking, and monitoring neurodegenerative diseases, particularly Alzheimer’s disease. While larger datasets often provide an advantage in deep learning-based segmentation, smaller da...

Full description

Saved in:
Bibliographic Details
Main Authors: Faizaan Fazal Khan, Jun-Hyung Kim, Chun-Su Park, Ji-In Kim, Goo-Rak Kwon
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11027081/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849772763718352896
author Faizaan Fazal Khan
Jun-Hyung Kim
Chun-Su Park
Ji-In Kim
Goo-Rak Kwon
author_facet Faizaan Fazal Khan
Jun-Hyung Kim
Chun-Su Park
Ji-In Kim
Goo-Rak Kwon
author_sort Faizaan Fazal Khan
collection DOAJ
description Hippocampus segmentation on MRI (magnetic resonance imaging) plays a vital role in detecting, diagnosing, tracking, and monitoring neurodegenerative diseases, particularly Alzheimer&#x2019;s disease. While larger datasets often provide an advantage in deep learning-based segmentation, smaller datasets pose unique challenges due to limited data variability and an increased risk of overfitting. This study addresses these challenges by developing a computationally efficient and accurate 3D U-Net model tailored for hippocampus segmentation. The proposed approach employs a preprocessing pipeline combining 3D Contrast Limited Adaptive Histogram Equalization (CLAHE) and Selective Coefficient-Enhanced 3D Wavelet Transform (SCE-3DWT), which enhances contrast and reduces noise for improved feature extraction. The experimental evaluation was conducted using the EADC-ADNI HarP dataset, comprising 135 hippocampal MRI scans with an input image size of <inline-formula> <tex-math notation="LaTeX">$64\times 64 \times 96$ </tex-math></inline-formula>. The model achieved a Dice coefficient of 0.8838 and a Jaccard Index of 0.7920, surpassing recent state-of-the-art methods. Comparative analysis highlights reduced Over-Segmentation Ratio (OSR = FP/(FP+TP), 0.0594) and Under-Segmentation Ratio (USR = FN/(FN+TP), 0.0569, reflecting its robustness and generalization. The lightweight architecture, designed with a maximum filter size of 512, operates efficiently without relying on transfer learning, making it accessible for broader applications. Future work will focus on integrating post processing techniques, leveraging larger and more diverse datasets, and exploring higher-resolution volumetric data to further improve segmentation accuracy and clinical utility. This study contributes to the advancement of medical image analysis, offering a resource-efficient framework for precise hippocampus segmentation, with potential implications for improved Alzheimer&#x2019;s disease management.
format Article
id doaj-art-b90fdbb4d5984edd8e801e319ead30ab
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-b90fdbb4d5984edd8e801e319ead30ab2025-08-20T03:02:14ZengIEEEIEEE Access2169-35362025-01-0113999239993810.1109/ACCESS.2025.357726611027081A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT TechniquesFaizaan Fazal Khan0https://orcid.org/0009-0000-8828-2695Jun-Hyung Kim1https://orcid.org/0000-0003-2231-5209Chun-Su Park2https://orcid.org/0000-0003-4250-2597Ji-In Kim3Goo-Rak Kwon4https://orcid.org/0000-0003-3486-8812Department of Information and Communication Engineering, Chosun University, Dong-gu, Gwangju, South KoreaDepartment of Information and Communication Engineering, Chosun University, Dong-gu, Gwangju, South KoreaDepartment of Computer Education, Sungkyunkwan University, Jongno-gu, Seoul, Republic of KoreaDepartment of Information and Communication Engineering, Chosun University, Dong-gu, Gwangju, South KoreaDepartment of Information and Communication Engineering, Chosun University, Dong-gu, Gwangju, South KoreaHippocampus segmentation on MRI (magnetic resonance imaging) plays a vital role in detecting, diagnosing, tracking, and monitoring neurodegenerative diseases, particularly Alzheimer&#x2019;s disease. While larger datasets often provide an advantage in deep learning-based segmentation, smaller datasets pose unique challenges due to limited data variability and an increased risk of overfitting. This study addresses these challenges by developing a computationally efficient and accurate 3D U-Net model tailored for hippocampus segmentation. The proposed approach employs a preprocessing pipeline combining 3D Contrast Limited Adaptive Histogram Equalization (CLAHE) and Selective Coefficient-Enhanced 3D Wavelet Transform (SCE-3DWT), which enhances contrast and reduces noise for improved feature extraction. The experimental evaluation was conducted using the EADC-ADNI HarP dataset, comprising 135 hippocampal MRI scans with an input image size of <inline-formula> <tex-math notation="LaTeX">$64\times 64 \times 96$ </tex-math></inline-formula>. The model achieved a Dice coefficient of 0.8838 and a Jaccard Index of 0.7920, surpassing recent state-of-the-art methods. Comparative analysis highlights reduced Over-Segmentation Ratio (OSR = FP/(FP+TP), 0.0594) and Under-Segmentation Ratio (USR = FN/(FN+TP), 0.0569, reflecting its robustness and generalization. The lightweight architecture, designed with a maximum filter size of 512, operates efficiently without relying on transfer learning, making it accessible for broader applications. Future work will focus on integrating post processing techniques, leveraging larger and more diverse datasets, and exploring higher-resolution volumetric data to further improve segmentation accuracy and clinical utility. This study contributes to the advancement of medical image analysis, offering a resource-efficient framework for precise hippocampus segmentation, with potential implications for improved Alzheimer&#x2019;s disease management.https://ieeexplore.ieee.org/document/11027081/Alzheimer’s disease3D contrast limited adaptive histogram equalization (CLAHE)EADC-ADNI Harmonized protocol (HarP) datasethippocampus segmentationselective coefficient-enhanced 3D wavelet transform for MRI segmentation (SCE-3DWT)3D U-Net
spellingShingle Faizaan Fazal Khan
Jun-Hyung Kim
Chun-Su Park
Ji-In Kim
Goo-Rak Kwon
A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
IEEE Access
Alzheimer’s disease
3D contrast limited adaptive histogram equalization (CLAHE)
EADC-ADNI Harmonized protocol (HarP) dataset
hippocampus segmentation
selective coefficient-enhanced 3D wavelet transform for MRI segmentation (SCE-3DWT)
3D U-Net
title A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
title_full A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
title_fullStr A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
title_full_unstemmed A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
title_short A Resource-Efficient 3D U-Net for Hippocampus Segmentation Using CLAHE and SCE-3DWT Techniques
title_sort resource efficient 3d u net for hippocampus segmentation using clahe and sce 3dwt techniques
topic Alzheimer’s disease
3D contrast limited adaptive histogram equalization (CLAHE)
EADC-ADNI Harmonized protocol (HarP) dataset
hippocampus segmentation
selective coefficient-enhanced 3D wavelet transform for MRI segmentation (SCE-3DWT)
3D U-Net
url https://ieeexplore.ieee.org/document/11027081/
work_keys_str_mv AT faizaanfazalkhan aresourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT junhyungkim aresourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT chunsupark aresourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT jiinkim aresourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT goorakkwon aresourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT faizaanfazalkhan resourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT junhyungkim resourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT chunsupark resourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT jiinkim resourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques
AT goorakkwon resourceefficient3dunetforhippocampussegmentationusingclaheandsce3dwttechniques