1
|
Liu X, Xing F, Gaggin HK, Kuo CCJ, El Fakhri G, Woo J. SUCCESSIVE SUBSPACE LEARNING FOR CARDIAC DISEASE CLASSIFICATION WITH TWO-PHASE DEFORMATION FIELDS FROM CINE MRI. PROCEEDINGS. IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING 2023; 2023:10.1109/isbi53787.2023.10230746. [PMID: 38031559 PMCID: PMC10686280 DOI: 10.1109/isbi53787.2023.10230746] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 12/01/2023]
Abstract
Cardiac cine magnetic resonance imaging (MRI) has been used to characterize cardiovascular diseases (CVD), often providing a noninvasive phenotyping tool. While recently flourished deep learning based approaches using cine MRI yield accurate characterization results, the performance is often degraded by small training samples. In addition, many deep learning models are deemed a "black box," for which models remain largely elusive in how models yield a prediction and how reliable they are. To alleviate this, this work proposes a lightweight successive subspace learning (SSL) framework for CVD classification, based on an interpretable feedforward design, in conjunction with a cardiac atlas. Specifically, our hierarchical SSL model is based on (i) neighborhood voxel expansion, (ii) unsupervised subspace approximation, (iii) supervised regression, and (iv) multi-level feature integration. In addition, using two-phase 3D deformation fields, including end-diastolic and end-systolic phases, derived between the atlas and individual subjects as input offers objective means of assessing CVD, even with small training samples. We evaluate our framework on the ACDC2017 database, comprising one healthy group and four disease groups. Compared with 3D CNN-based approaches, our framework achieves superior classification performance with 140× fewer parameters, which supports its potential value in clinical use.
Collapse
Affiliation(s)
- Xiaofeng Liu
- Dept. of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - Fangxu Xing
- Dept. of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - Hanna K Gaggin
- Dept. of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - C-C Jay Kuo
- Dept. of ECE, University of Southern California, Los Angeles, CA, USA
| | - Georges El Fakhri
- Dept. of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - Jonghye Woo
- Dept. of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| |
Collapse
|
2
|
Liu X, Yoo C, Xing F, Kuo CCJ, El Fakhri G, Kang JW, Woo J. Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation. Front Neurosci 2022; 16:837646. [PMID: 35720708 PMCID: PMC9201342 DOI: 10.3389/fnins.2022.837646] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/16/2021] [Accepted: 02/22/2022] [Indexed: 02/03/2023] Open
Abstract
Unsupervised domain adaptation (UDA) is an emerging technique that enables the transfer of domain knowledge learned from a labeled source domain to unlabeled target domains, providing a way of coping with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of privacy concerns about potential leaks in sensitive information contained in patient data, it is often challenging to share the data and labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on a few datasets and deep learning backbones, demonstrating the potential for our framework to be applied in challenging yet realistic clinical settings.
Collapse
Affiliation(s)
- Xiaofeng Liu
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States
| | - Chaehwa Yoo
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States,Department of Electronic and Electrical Engineering and Graduate Program in Smart Factory, Ewha Womans University, Seoul, South Korea
| | - Fangxu Xing
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States
| | - C.-C. Jay Kuo
- Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, United States
| | - Georges El Fakhri
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States
| | - Je-Won Kang
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States,Department of Electronic and Electrical Engineering and Graduate Program in Smart Factory, Ewha Womans University, Seoul, South Korea
| | - Jonghye Woo
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States,*Correspondence: Jonghye Woo
| |
Collapse
|
3
|
Liu X, Yoo C, Xing F, Kuo CCJ, El Fakhri G, Kang JW, Woo J. Unsupervised Domain Adaptation for Segmentation with Black-box Source Model. PROCEEDINGS OF SPIE--THE INTERNATIONAL SOCIETY FOR OPTICAL ENGINEERING 2022; 12032:1203210. [PMID: 35983176 PMCID: PMC9385170 DOI: 10.1117/12.2607895] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/15/2023]
Abstract
Unsupervised domain adaptation (UDA) has been widely used to transfer knowledge from a labeled source domain to an unlabeled target domain to counter the difficulty of labeling in a new domain. The training of conventional solutions usually relies on the existence of both source and target domain data. However, privacy of the large-scale and well-labeled data in the source domain and trained model parameters can become the major concern of cross center/domain collaborations. In this work, to address this, we propose a practical solution to UDA for segmentation with a black-box segmentation model trained in the source domain only, rather than original source data or a white-box source model. Specifically, we resort to a knowledge distillation scheme with exponential mixup decay (EMD) to gradually learn target-specific representations. In addition, unsupervised entropy minimization is further applied to regularization of the target domain confidence. We evaluated our framework on the BraTS 2018 database, achieving performance on par with white-box source model adaptation approaches.
Collapse
Affiliation(s)
- Xiaofeng Liu
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - Chaehwa Yoo
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
- Dept. of Electronic and Electrical Engineering and Graduate Program in Smart Factory, Ewha Womans University, Seoul, South Korea
| | - Fangxu Xing
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - C-C Jay Kuo
- Dept. of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, USA
| | - Georges El Fakhri
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| | - Je-Won Kang
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
- Dept. of Electronic and Electrical Engineering and Graduate Program in Smart Factory, Ewha Womans University, Seoul, South Korea
| | - Jonghye Woo
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
| |
Collapse
|
4
|
Liu X, Xing F, Prince JL, Stone M, El Fakhri G, Woo J. Structure-aware Unsupervised Tagged-to-Cine MRI Synthesis with Self Disentanglement. PROCEEDINGS OF SPIE--THE INTERNATIONAL SOCIETY FOR OPTICAL ENGINEERING 2022; 12032:120321Q. [PMID: 36203947 PMCID: PMC9533681 DOI: 10.1117/12.2610655] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/16/2023]
Abstract
Cycle reconstruction regularized adversarial training-e.g., CycleGAN, DiscoGAN, and DualGAN-has been widely used for image style transfer with unpaired training data. Several recent works, however, have shown that local distortions are frequent, and structural consistency cannot be guaranteed. Targeting this issue, prior works usually relied on additional segmentation or consistent feature extraction steps that are task-specific. To counter this, this work aims to learn a general add-on structural feature extractor, by explicitly enforcing the structural alignment between an input and its synthesized image. Specifically, we propose a novel input-output image patches self-training scheme to achieve a disentanglement of underlying anatomical structures and imaging modalities. The translator and structure encoder are updated, following an alternating training protocol. In addition, the information w.r.t. imaging modality can be eliminated with an asymmetric adversarial game. We train, validate, and test our network on 1,768, 416, and 1,560 unpaired subject-independent slices of tagged and cine magnetic resonance imaging from a total of twenty healthy subjects, respectively, demonstrating superior performance over competing methods.
Collapse
Affiliation(s)
- Xiaofeng Liu
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114 USA
| | - Fangxu Xing
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114 USA
| | - Jerry L Prince
- Deportment of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, MD 21218 USA
| | - Maureen Stone
- Deportment of Neural and Pain Sciences, University of Maryland School of Dentistry, Baltimore, MD 21201 USA
| | - Georges El Fakhri
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114 USA
| | - Jonghye Woo
- Gordon Center for Medical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114 USA
| |
Collapse
|