3D MRI Brain Tumor Segmentation Using Autoencoder Regularization

Myronenko, Andriy

doi:10.1007/978-3-030-11726-9_28

Cited by 786 publications

(588 citation statements)

References 23 publications

Supporting

Mentioning

517

Contrasting

Unclassified

Order By: Relevance

“…Therefore, it improves the segmentation quality around the edges, which leads to overall better segmentation performance. Figure 2 illustrates how the addition of the EG-CNN to a standalone Seg-Net [16] improves the quality of segmentation. The quality of the predicted edges also validates the effectiveness of our proposed edgeaware loss function, since the boundaries are crisp and avoid the thickening effect around edges.…”

Section: Resultsmentioning

confidence: 99%

Edge-Gated CNNs for Volumetric Semantic Segmentation of Medical Images

Hatamizadeh

Terzopoulos

Myronenko

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Textures and edges contribute different information to image recognition. Edges and boundaries encode shape information, while textures manifest the appearance of regions. Despite the success of Convolutional Neural Networks (CNNs) in computer vision and medical image analysis applications, predominantly only texture abstractions are learned, which often leads to imprecise boundary delineations. In medical imaging, expert manual segmentation often relies on organ boundaries; for example, to manually segment a liver, a medical practitioner usually identifies edges first and subsequently fills in the segmentation mask. Motivated by these observations, we propose a plug-and-play module, dubbed Edge-Gated CNNs (EG-CNNs), that can be used with existing encoder-decoder architectures to process both edge and texture information. The EG-CNN learns to emphasize the edges in the encoder, to predict crisp boundaries by an auxiliary edge supervision, and to fuse its output with the original CNN output. We evaluate the effectiveness of the EG-CNN with various mainstream CNNs on two publicly available datasets, BraTS 19 and KiTS 19 for brain tumor and kidney semantic segmentation. We demonstrate how the addition of EG-CNN consistently improves segmentation accuracy and generalization performance. c 2020 A. Hatamizadeh, D. Terzopoulos & A. Myronenko.

show abstract

Section: Resultsmentioning

confidence: 99%

Edge-Gated CNNs for Volumetric Semantic Segmentation of Medical Images

Hatamizadeh

Terzopoulos

Myronenko

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In addition, we added volume and surface area features of each tumor component (Feng, Tustison et al 2018) and age. We performed feature selection based on SelectKBest features using the sklearn package (Pedregosa, Varoquaux et al 2011, Buitinck, Louppe et al 2013) which resulted in a reduced set of 25 features. We trained an XGBoost (XGB) model (Chen and Guestrin 2016) to classify the subjects into low (less than 300 days), medium (between 300 to 450 days) and long survivors (greater than 450 days).…”

Section: Ensemble Methodsologymentioning

confidence: 99%

“…Havaei et al developed a multiresolution cascaded CNN architecture with two pathways, each of which takes different 2D patch sizes with four MR sequences as channels (Havaei, Davy et al 2017). The BRATS 2018 top performer developed a 3D decoder encoder style CNN architecture with inter-level skip connections to segment the tumor (Myronenko 2018). In addition to the decoder part, a Variation Autoencoder (VAE) was included to add reconstruction loss to the model.…”

mentioning

confidence: 99%

Multidimensional and Multiresolution Ensemble Networks for Brain Tumor Segmentation

Murugesan

Nalawade

Ganesh

et al. 2019

Preprint

View full text Add to dashboard Cite

In this work, we developed multiple 2D and 3D segmentation models with multiresolution input to segment brain tumor components, and then ensembled them to obtain robust segmentation maps. This reduced overfitting and resulted in a more generalized model. Multiparametric MR images of 335 subjects from BRATS 2019 challenge were used for training the models. Further, we tested a classical machine learning algorithm (xgboost) with features extracted from the segmentation maps to classify subject survival range. Preliminary results on the BRATS 2019 validation dataset demonstrasted this method can achieve excellent performance with DICE scores of 0.898, 0.784, 0.779 for whole tumor, tumor core and enhancing tumor respectively and accuracy 34.5 % for survuval prediction. * -Equal Contribution

show abstract

“…Therefore, automated and robust methods for providing a segmentation of the cardiac anatomy around the left ventricle (LV) are needed to support the analysis of myocardial infarction. Modern semantic segmentation methods utilizing deep learning have significantly improved the performance in various medical imaging applications [3,4,5,6]. At the same time, deep learning methods typically require large amounts of annotated data in order to train sufficiently robust and accurate models depending on the difficulty of the task.…”

Section: Introductionmentioning

confidence: 99%

Cardiac Segmentation of LGE MRI with Noisy Labels

Roth¹,

Zhu²,

Yang³

et al. 2020

Statistical Atlases and Computational Models of the Heart. Multi-Sequence CMR Segmentation, CRT-EPiggy and LV Full Quantificati

View full text Add to dashboard Cite

In this work, we attempt the segmentation of cardiac structures in late gadolinium-enhanced (LGE) magnetic resonance images (MRI) using only minimal supervision in a two-step approach. In the first step, we register a small set of five LGE cardiac magnetic resonance (CMR) images with ground truth labels to a set of 40 target LGE CMR images without annotation. Each manually annotated ground truth provides labels of the myocardium and the left ventricle (LV) and right ventricle (RV) cavities, which are used as atlases. After multi-atlas label fusion by majority voting, we possess noisy labels for each of the targeted LGE images. A second set of manual labels exists for 30 patients of the target LGE CMR images, but are annotated on different MRI sequences (bSSFP and T2-weighted). Again, we use multi-atlas label fusion with a consistency constraint to further refine our noisy labels if additional annotations in other modalities are available for a given patient. In the second step, we train a deep convolutional network for semantic segmentation on the target data while using data augmentation techniques to avoid over-fitting to the noisy labels. After inference and simple post-processing, we achieve our final segmentation for the targeted LGE CMR images, resulting in an average Dice of 0.890, 0.780, and 0.844 for LV cavity, LV myocardium, and RV cavity, respectively.

show abstract

3D MRI Brain Tumor Segmentation Using Autoencoder Regularization

Cited by 786 publications

References 23 publications

Edge-Gated CNNs for Volumetric Semantic Segmentation of Medical Images

Edge-Gated CNNs for Volumetric Semantic Segmentation of Medical Images

Multidimensional and Multiresolution Ensemble Networks for Brain Tumor Segmentation

Cardiac Segmentation of LGE MRI with Noisy Labels

Contact Info

Product

Resources

About