Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging
<p>Cardiovascular magnetic resonance (CMR) imaging is a powerful tool for research and clinical applications. To extract useful clinical information from the acquired CMR images, time-consuming and laborious manual delineation of cardiovascular structures is currently required. Despite promisi...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2020
|
| Subjects: |
| _version_ | 1797113344669253632 |
|---|---|
| author | Hann, E |
| author2 | Piechnik, S |
| author_facet | Piechnik, S Hann, E |
| author_sort | Hann, E |
| collection | OXFORD |
| description | <p>Cardiovascular magnetic resonance (CMR) imaging is a powerful tool for research and clinical applications. To extract useful clinical information from the acquired CMR images, time-consuming and laborious manual delineation of cardiovascular structures is currently required. Despite promising overall performance across medical imaging applications, the current state-of-the-art automated image segmentation methods still fail in some cases, potentially jeopardising the reliability of clinical diagnosis. Thus, it is important to develop not only automation of image segmentation but also quality control of segmentation, to empower efficient and reliable CMR image data analysis.</p>
<p>To address both segmentation and quality control problems, I have developed a novel quality control-driven (QCD) framework in this thesis. Extending upon deep ensemble learning, the framework utilises multiple convolutional neural network-based models to generate segmentation candidates, the agreement of which is exploited via additional regression models to predict segmentation quality measured by Dice similarity coefficient (DSC). The DSC prediction not only provides a quality estimate but also enables a novel approach to select a final, most optimal segmentation on-the-fly from multiple candidates, improving segmentation robustness. Following the DSC prediction, a segmentation quality classification scheme is implemented to alert human operators only when manual intervention is recommended, intended for more efficient allocation of time and labour resources for large-scale image processing pipelines.</p>
<p>Through both quantitative and qualitative evaluation, the QCD framework has demonstrated excellent performance in both segmentation and quality control. More importantly, the framework has been successfully applied across CMR imaging techniques, anatomical structures, and large-scale datasets acquired at different sites, with high adaptability and generalisability. The QCD framework could pave the way towards large-scale automated imaging data analysis pipelines, with both efficiency and reliability, in real-world clinical applications.</p> |
| first_indexed | 2024-03-07T06:11:19Z |
| format | Thesis |
| id | oxford-uuid:ef9a524c-07b3-4b7d-936d-da862f81f15a |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-04-23T08:27:20Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | oxford-uuid:ef9a524c-07b3-4b7d-936d-da862f81f15a2024-04-22T09:26:18ZDeep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imagingThesishttp://purl.org/coar/resource_type/c_db06uuid:ef9a524c-07b3-4b7d-936d-da862f81f15aSupervised learning (Machine learning)Magnetic resonance imagingMachine learningEnsemble learning (Machine learning)Diagnostic imagingImage processingImage segmentationArtificial intelligenceNeural networks (Computer science)EnglishHyrax Deposit2020Hann, EPiechnik, SFerreira, VNeubauer, SGrau Colomer, VFontana, M<p>Cardiovascular magnetic resonance (CMR) imaging is a powerful tool for research and clinical applications. To extract useful clinical information from the acquired CMR images, time-consuming and laborious manual delineation of cardiovascular structures is currently required. Despite promising overall performance across medical imaging applications, the current state-of-the-art automated image segmentation methods still fail in some cases, potentially jeopardising the reliability of clinical diagnosis. Thus, it is important to develop not only automation of image segmentation but also quality control of segmentation, to empower efficient and reliable CMR image data analysis.</p> <p>To address both segmentation and quality control problems, I have developed a novel quality control-driven (QCD) framework in this thesis. Extending upon deep ensemble learning, the framework utilises multiple convolutional neural network-based models to generate segmentation candidates, the agreement of which is exploited via additional regression models to predict segmentation quality measured by Dice similarity coefficient (DSC). The DSC prediction not only provides a quality estimate but also enables a novel approach to select a final, most optimal segmentation on-the-fly from multiple candidates, improving segmentation robustness. Following the DSC prediction, a segmentation quality classification scheme is implemented to alert human operators only when manual intervention is recommended, intended for more efficient allocation of time and labour resources for large-scale image processing pipelines.</p> <p>Through both quantitative and qualitative evaluation, the QCD framework has demonstrated excellent performance in both segmentation and quality control. More importantly, the framework has been successfully applied across CMR imaging techniques, anatomical structures, and large-scale datasets acquired at different sites, with high adaptability and generalisability. The QCD framework could pave the way towards large-scale automated imaging data analysis pipelines, with both efficiency and reliability, in real-world clinical applications.</p> |
| spellingShingle | Supervised learning (Machine learning) Magnetic resonance imaging Machine learning Ensemble learning (Machine learning) Diagnostic imaging Image processing Image segmentation Artificial intelligence Neural networks (Computer science) Hann, E Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title | Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title_full | Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title_fullStr | Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title_full_unstemmed | Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title_short | Deep ensemble learning-based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| title_sort | deep ensemble learning based quality control for automatic segmentation in cardiovascular magnetic resonance imaging |
| topic | Supervised learning (Machine learning) Magnetic resonance imaging Machine learning Ensemble learning (Machine learning) Diagnostic imaging Image processing Image segmentation Artificial intelligence Neural networks (Computer science) |
| work_keys_str_mv | AT hanne deepensemblelearningbasedqualitycontrolforautomaticsegmentationincardiovascularmagneticresonanceimaging |