Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images

Abstract Multi‐organ segmentation from abdominal CT scans is crucial for various medical examinations and diagnoses. Despite the remarkable achievements of existing deep‐learning‐based methods, accurately segmenting small organs remains challenging due to their small size and low contrast. This arti...

Full description

Saved in:
Bibliographic Details
Main Authors: Tao Liu, Xukun Zhang, Zhongwei Yang, Minghao Han, Haopeng Kuang, Shuwei Ma, Le Wang, Xiaoying Wang, Lihua Zhang
Format: Article
Language:English
Published: Wiley 2024-11-01
Series:IET Image Processing
Subjects:
Online Access:https://doi.org/10.1049/ipr2.13221
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850191371325931520
author Tao Liu
Xukun Zhang
Zhongwei Yang
Minghao Han
Haopeng Kuang
Shuwei Ma
Le Wang
Xiaoying Wang
Lihua Zhang
author_facet Tao Liu
Xukun Zhang
Zhongwei Yang
Minghao Han
Haopeng Kuang
Shuwei Ma
Le Wang
Xiaoying Wang
Lihua Zhang
author_sort Tao Liu
collection DOAJ
description Abstract Multi‐organ segmentation from abdominal CT scans is crucial for various medical examinations and diagnoses. Despite the remarkable achievements of existing deep‐learning‐based methods, accurately segmenting small organs remains challenging due to their small size and low contrast. This article introduces a novel knowledge‐guided cascaded framework that utilizes two types of knowledge—image intrinsic (anatomy) and clinical expertise (radiology)—to improve the segmentation accuracy of small abdominal organs. Specifically, based on the anatomical similarities in abdominal CT scans, the approach employs entropy‐based registration techniques to map high‐quality segmentation results onto inaccurate results from the first stage, thereby guiding precise localization of small organs. Additionally, inspired by the practice of annotating images from multiple perspectives by radiologists, novel Multi‐View Fusion Convolution (MVFC) operator is developed, which can extract and adaptively fuse features from various directions of CT images to refine segmentation of small organs effectively. Simultaneously, the MVFC operator offers a seamless alternative to conventional convolutions within diverse model architectures. Extensive experiments on the Abdominal Multi‐Organ Segmentation (AMOS) dataset demonstrate the superiority of the method, setting a new benchmark in the segmentation of small organs.
format Article
id doaj-art-a5acf9ebfba24090837fb882fceae63a
institution OA Journals
issn 1751-9659
1751-9667
language English
publishDate 2024-11-01
publisher Wiley
record_format Article
series IET Image Processing
spelling doaj-art-a5acf9ebfba24090837fb882fceae63a2025-08-20T02:14:56ZengWileyIET Image Processing1751-96591751-96672024-11-0118133935394910.1049/ipr2.13221Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT imagesTao Liu0Xukun Zhang1Zhongwei Yang2Minghao Han3Haopeng Kuang4Shuwei Ma5Le Wang6Xiaoying Wang7Lihua Zhang8Academy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaZhongshan Hospital Fudan University Shanghai ChinaAcademy for Engineering and Technology Fudan University Shanghai ChinaAbstract Multi‐organ segmentation from abdominal CT scans is crucial for various medical examinations and diagnoses. Despite the remarkable achievements of existing deep‐learning‐based methods, accurately segmenting small organs remains challenging due to their small size and low contrast. This article introduces a novel knowledge‐guided cascaded framework that utilizes two types of knowledge—image intrinsic (anatomy) and clinical expertise (radiology)—to improve the segmentation accuracy of small abdominal organs. Specifically, based on the anatomical similarities in abdominal CT scans, the approach employs entropy‐based registration techniques to map high‐quality segmentation results onto inaccurate results from the first stage, thereby guiding precise localization of small organs. Additionally, inspired by the practice of annotating images from multiple perspectives by radiologists, novel Multi‐View Fusion Convolution (MVFC) operator is developed, which can extract and adaptively fuse features from various directions of CT images to refine segmentation of small organs effectively. Simultaneously, the MVFC operator offers a seamless alternative to conventional convolutions within diverse model architectures. Extensive experiments on the Abdominal Multi‐Organ Segmentation (AMOS) dataset demonstrate the superiority of the method, setting a new benchmark in the segmentation of small organs.https://doi.org/10.1049/ipr2.13221computerised tomographylearning (artificial intelligence)medical image processing
spellingShingle Tao Liu
Xukun Zhang
Zhongwei Yang
Minghao Han
Haopeng Kuang
Shuwei Ma
Le Wang
Xiaoying Wang
Lihua Zhang
Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
IET Image Processing
computerised tomography
learning (artificial intelligence)
medical image processing
title Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
title_full Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
title_fullStr Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
title_full_unstemmed Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
title_short Dual knowledge‐guided two‐stage model for precise small organ segmentation in abdominal CT images
title_sort dual knowledge guided two stage model for precise small organ segmentation in abdominal ct images
topic computerised tomography
learning (artificial intelligence)
medical image processing
url https://doi.org/10.1049/ipr2.13221
work_keys_str_mv AT taoliu dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT xukunzhang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT zhongweiyang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT minghaohan dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT haopengkuang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT shuweima dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT lewang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT xiaoyingwang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages
AT lihuazhang dualknowledgeguidedtwostagemodelforprecisesmallorgansegmentationinabdominalctimages