Adaptive Hierarchical Dual Consistency for Semi-Supervised Left Atrium Segmentation on Cross-Domain Data

Chen, Jun; Zhang, Heye; Mohiaddin, Raad H.; Wong, Tom; Firmin, David; Keegan, Jennifer; Yang, Guang

doi:10.1109/tmi.2021.3113678

Cited by 34 publications

(10 citation statements)

References 42 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Differently from classical domain adaptation considering only cross-modality domain shift for segmenting the same set of structures [6], [32], we deal with a more difficult scenario where an existing dataset with similar anatomical structures is used to assist model training in the target domain. Here, the "similar anatomical structure" requirement means there is a shape similarity between the two domains, i.e., the shapes have similar topologies but may be different in scales.…”

Section: Discussionmentioning

confidence: 99%

“…The first is fully-supervised DA, where fine-tuning is the most representative technique for adapting a trained model to the target domain with full annotations [30], [31]. The second is weakly or semi-supervised DA where only coarse or partial annotations in the target domain are used for model adaptation [32]. For example, Dorent et al [33] employed scribbles in the target domain to perform model adaptation for vestibular schwannoma segmentation.…”

Section: B Transfer Learning and Domain Adaptationmentioning

confidence: 99%

See 1 more Smart Citation

Contrastive Semi-supervised Learning for Domain Adaptive Segmentation Across Similar Anatomical Structures

Gu¹,

Zhang²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance for medical image segmentation, yet need plenty of manual annotations for training. Semi-Supervised Learning (SSL) methods are promising to reduce the requirement of annotations, but their performance is still limited when the dataset size and the number of annotated images are small. Leveraging existing annotated datasets with similar anatomical structures to assist training has a potential for improving the model's performance. However, it is further challenged by the cross-anatomy domain shift due to the different appearance and even imaging modalities from the target structure. To solve this problem, we propose Contrastive Semisupervised learning for Cross Anatomy Domain Adaptation (CS-CADA) that adapts a model to segment similar structures in a target domain, which requires only limited annotations in the target domain by leveraging a set of existing annotated images of similar structures in a source domain. We use Domain-Specific Batch Normalization (DSBN) to individually normalize feature maps for the two anatomical domains, and propose a cross-domain contrastive learning strategy to encourage extracting domain invariant features. They are integrated into a Self-Ensembling Mean-Teacher (SE-MT) framework to exploit unlabeled target domain images with a prediction consistency constraint. Extensive experiments show that our CS-CADA is able to solve the challenging cross-anatomy domain shift problem, achieving accurate segmentation of coronary arteries in X-ray images with the help of retinal vessel images and cardiac MR images with the help of fundus images, respectively, given only a small number of annotations in the target domain.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: B Transfer Learning and Domain Adaptationmentioning

confidence: 99%

Contrastive Semi-supervised Learning for Domain Adaptive Segmentation Across Similar Anatomical Structures

Gu¹,

Zhang²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Meanwhile, to remain discriminative for the segmentation of labeled and unlabeled data, further segmentation supervision is obtained through comparing the non-local semantic relation matrix in feature maps from the ground truth label mask and the student inputs. Another work in [120] propose adaptive hierarchical dual consistency to use the dataset from different centers, which learns mapping networks adversarially to align the distributions and extend consistency learning into intra-and inter-consistency in cross-domain segmentation. Another idea for using data form multi centers is through meta-learning.…”

Section: Other Semi-supervised Medical Image Segmentation Methodsmentioning

confidence: 99%

Learning with Limited Annotations: A Survey on Deep Semi-Supervised Learning for Medical Image Segmentation

Jiao¹,

Zhang²,

Liu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Medical image segmentation is a fundamental and critical step in many image-guided clinical approaches. Recent success of deep learning-based segmentation methods usually relies on a large amount of labeled data, which is particularly difficult and costly to obtain especially in the medical imaging domain where only experts can provide reliable and accurate annotations. Semi-supervised learning has emerged as an appealing strategy and been widely applied to medical image segmentation tasks to train deep models with limited annotations. In this paper, we present a comprehensive review of recently proposed semi-supervised learning methods for medical image segmentation and summarized both the technical novelties and empirical results. Furthermore, we analyze and discuss the limitations and several unsolved problems of existing approaches. We hope this review could inspire the research community to explore solutions for this challenge and further promote the developments in medical image segmentation field.

show abstract

“…To measure the biological accuracy of the CNN predictions, we used the error in the LA diameter and volume. These are important biomarkers which have been shown to provide reliable information during clinical diagnosis and treatment stratification of atrial fibrillation ( Zhuang et al, 2011 ; Njoku et al, 2018 ; Chen et al, 2022 ). The LA diameter was defined as

for a 2D slice of the 3D LA geometry with the maximum 2D width to obtain the overall maximum LA diameter, M , with dimensions I × J , where J was the anterior-posterior axis of the LA chamber.…”

Section: Methodsmentioning

confidence: 99%

Automatic 3D Surface Reconstruction of the Left Atrium From Clinically Mapped Point Clouds Using Convolutional Neural Networks

et al. 2022

View full text Add to dashboard Cite

Point clouds are a widely used format for storing information in a memory-efficient and easily manipulatable representation. However, research in the application of point cloud mapping and subsequent organ reconstruction with deep learning, is limited. In particular, current methods for left atrium (LA) visualization using point clouds recorded from clinical mapping during cardiac ablation are proprietary and remain difficult to validate. Many clinics rely on additional imaging such as MRIs/CTs to improve the accuracy of LA mapping. In this study, for the first time, we proposed a novel deep learning framework for the automatic 3D surface reconstruction of the LA directly from point clouds acquired via widely used clinical mapping systems. The backbone of our framework consists of a 30-layer 3D fully convolutional neural network (CNN). The architecture contains skip connections that perform multi-resolution processing to maximize information extraction from the point clouds and ensure a high-resolution prediction by combining features at different receptive levels. We used large kernels with increased receptive fields to address the sparsity of the point clouds. Residual blocks and activation normalization were further implemented to improve the feature learning on sparse inputs. By utilizing a light-weight design with low-depth layers, our CNN took approximately 10 s per patient. Independent testing on two cross-modality clinical datasets showed excellent dice scores of 93% and surface-to-surface distances below 1 pixel. Overall, our study may provide a more efficient, cost-effective 3D LA reconstruction approach during ablation procedures, and potentially lead to improved treatment of cardiac diseases.

show abstract

Adaptive Hierarchical Dual Consistency for Semi-Supervised Left Atrium Segmentation on Cross-Domain Data

Cited by 34 publications

References 42 publications

Contrastive Semi-supervised Learning for Domain Adaptive Segmentation Across Similar Anatomical Structures

Contrastive Semi-supervised Learning for Domain Adaptive Segmentation Across Similar Anatomical Structures

Learning with Limited Annotations: A Survey on Deep Semi-Supervised Learning for Medical Image Segmentation

Automatic 3D Surface Reconstruction of the Left Atrium From Clinically Mapped Point Clouds Using Convolutional Neural Networks

Contact Info

Product

Resources

About