Filters








1,779 Hits in 8.2 sec

Generalisable Cardiac Structure Segmentation via Attentional and Stacked Image Adaptation [article]

Hongwei Li, Jianguo Zhang, Bjoern Menze
2020 arXiv   pre-print
A stack of data augmentation techniques was further used to simulate real-world transformation to boost the segmentation performance for unseen domains.We achieved an average Dice score of 90.3% for the  ...  A generative adversarial networks with an attention loss was proposed to translate the images from existing source domains to a target domain, thus to generate good-quality synthetic cardiac structure  ...  We proposed a sequence of k stacked image transformations f (·) to simulate image distributions for unseen domains.  ... 
arXiv:2008.01216v2 fatcat:csiaw3elcbdylhgpopdot5scxm

Causality-inspired Single-source Domain Generalization for Medical Image Segmentation [article]

Cheng Ouyang, Chen Chen, Surui Li, Zeju Li, Chen Qin, Wenjia Bai, Daniel Rueckert
2021 arXiv   pre-print
Deep learning models usually suffer from domain shift issues, where models trained on one source domain do not generalize well to other unseen domains.  ...  In this work, we investigate the single-source domain generalization problem: training a deep network that is robust to unseen domains, under the condition that training data is only available from one  ...  From a methodolog- for medical image segmentation to unseen domains via deep stacked ical perspective, while previous multi-source domain gener- transformation  ... 
arXiv:2111.12525v4 fatcat:e4ht2fhmijherf6jxo3fgmbgkm

Domain and Content Adaptive Convolution for Domain Generalization in Medical Image Segmentation [article]

Shishuai Hu, Zehui Liao, Jianpeng Zhang, Yong Xia
2021 arXiv   pre-print
In this paper, we propose a multi-source domain generalization model, namely domain and content adaptive convolution (DCAC), for medical image segmentation.  ...  The domain gap caused mainly by variable medical image quality renders a major obstacle on the path between training a segmentation model in the lab and applying the trained model to unseen clinical data  ...  Augmentation-based methods, such as the deep stacked transformation , simulate the distribution of target domain data by augmenting the source domain data.  ... 
arXiv:2109.05676v1 fatcat:tqufa3orfbdxtjeo7762ij6kyy

Model-Based and Data-Driven Strategies in Medical Image Computing [article]

Daniel Rueckert, Julia A. Schnabel
2019 arXiv   pre-print
We also discuss some of the open challenges for data-driven approaches, e.g. generalization to new unseen data (e.g. transfer learning), robustness to adversarial attacks and interpretability.  ...  These approaches learn statistical models directly from labelled or unlabeled image data and have been shown to be very powerful for extracting clinically useful information from medical imaging.  ...  IEH Award (NS/A000025/1) and the Innovate UK London Medical Imaging and AI Centre for Value Based Healthcare.  ... 
arXiv:1909.10391v3 fatcat:m3kpdg2sf5ai7cfjokhy7so2iy

FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space [article]

Quande Liu, Cheng Chen, Jing Qin, Qi Dou, Pheng-Ann Heng
2021 arXiv   pre-print
directly generalize to unseen target domains.  ...  the challenges of model generalization in medical image segmentation scenario.  ...  Introduction Data collaboration across multiple medical institutions is increasingly desired to build accurate and robust data-driven deep networks for medical image segmentation [7, 18, 50] .  ... 
arXiv:2103.06030v1 fatcat:cr4tiu6e4jgnfnt6wlt3scknmu

Biomechanics-informed Neural Networks for Myocardial Motion Tracking in MRI [article]

Chen Qin, Shuo Wang, Chen Chen, Huaqi Qiu, Wenjia Bai, Daniel Rueckert
2020 arXiv   pre-print
The method has also been shown to have better generalisability to unseen domains compared with commonly used L2 regularisation schemes.  ...  The learnt VAE regulariser then can be coupled with any deep learning based registration network to regularise the solution space to be biomechanically plausible.  ...  The authors wish to thank all UK Biobank participants and staff.  ... 
arXiv:2006.04725v3 fatcat:2bqejh4xzbd5rgrkbaumonevre

Learning Interpretable Anatomical Features Through Deep Generative Models: Application to Cardiac Remodeling [chapter]

Carlo Biffi, Ozan Oktay, Giacomo Tarroni, Wenjia Bai, Antonio De Marvao, Georgia Doumou, Martin Rajchl, Reem Bedair, Sanjay Prasad, Stuart Cook, Declan O'Regan, Daniel Rueckert
2018 Lecture Notes in Computer Science  
We believe that the proposed deep learning approach is a promising step towards the development of interpretable classifiers for the medical imaging domain, which may help clinicians to improve diagnostic  ...  Deep learning approaches have recently achieved success for tasks such as classification or segmentation of medical images, but lack interpretability in the feature extraction and decision processes, limiting  ...  The proposed approach is a promising step towards the development of interpretable deep learning classifiers for the medical imaging domain, which may assist clinicians to improve diagnosis and provide  ... 
doi:10.1007/978-3-030-00934-2_52 fatcat:fygzl4mxxffwtbcwcp3smsd2qi

DoFE: Domain-oriented Feature Embedding for Generalizable Fundus Image Segmentation on Unseen Datasets

Shujun Wang, Lequan Yu, Kang Li, Xin Yang, Chi-Wing Fu, Pheng-Ann Heng
2020 IEEE Transactions on Medical Imaging  
However, in clinical practice, medical images often exhibit variations in appearance for various reasons, e.g., different scanner vendors and image quality.  ...  These distribution discrepancies could lead the deep networks to over-fit on the training datasets and lack generalization ability on the unseen test datasets.  ...  [12] proposed a deep-stacked transformations approach for medical image segmentation by combining different kinds of data augmentation. Meanwhile, Chen et al.  ... 
doi:10.1109/tmi.2020.3015224 pmid:32776876 fatcat:4qpzzut5c5gnpj7okrhdoo4fmi

Joint Motion Correction and Super Resolution for Cardiac Segmentation via Latent Optimisation [article]

Shuo Wang, Chen Qin, Nicolo Savioli, Chen Chen, Declan O'Regan, Stuart Cook, Yike Guo, Daniel Rueckert, Wenjia Bai
2021 arXiv   pre-print
This alleviates the need of paired low-resolution and high-resolution images for supervised learning.  ...  However, due to the limit of acquisition duration and respiratory/cardiac motion, stacks of multi-slice 2D images are acquired in clinical routine.  ...  Supervised learning for super-resolution: Since the first attempt using CNN for single image super-resolution (SR) [6] , deep learning-based SR algorithms have successfully transformed the state-of-the-art  ... 
arXiv:2107.03887v1 fatcat:bkhxamheijg7tb7mxwmecq6pmu

Domain generalization in deep learning-based mass detection in mammography: A large-scale multi-center study [article]

Lidia Garrucho, Kaisar Kushibar, Socayna Jouide, Oliver Diaz, Laura Igual, Karim Lekadir
2022 arXiv   pre-print
Ultimately, this comprehensive study provides key insights and best practices for future research on domain generalization in deep learning-based breast cancer detection.  ...  To this end, we compare the performance of eight state-of-the-art detection methods, including Transformer-based models, trained in a single domain and tested in five unseen domains.  ...  We are also thanks to Volpara Health (Dr Melissa Hill) for agreeing to share the breast density information available of the OPTIMAM subset used.  ... 
arXiv:2201.11620v1 fatcat:twoja4j2d5cfnhrpjqmo36j3ky

Deep Learning in Medical Image Registration: A Survey [article]

Grant Haskins, Uwe Kruger, Pingkun Yan
2019 arXiv   pre-print
Since the beginning of the recent deep learning renaissance, the medical imaging research community has developed deep learning based approaches and achieved the state-of-the-art in many applications,  ...  The rapid adoption of deep learning for image registration applications over the past few years necessitates a comprehensive summary and outlook, which is the main scope of this survey.  ...  First, deep learning based medical image registration seems to be following the observed trend for the general application of deep learning to medical image analysis.  ... 
arXiv:1903.02026v1 fatcat:6ulnzrbj6rb55eydtkgygg5r6u

PnP-AdaNet: Plug-and-Play Adversarial Domain Adaptation Network at Unpaired Cross-modality Cardiac Segmentation

Qi Dou, Cheng Ouyang, Cheng Chen, Hao Chen, Ben Glocker, Xiahai Zhuang, Pheng Ann Heng
2019 IEEE Access  
Our code is publically available at https://github.com/carrenD/Medical-Cross-Modality-Domain-Adaptation INDEX TERMS Domain adaptation, adversarial learning, cardiac segmentation, medical imaging.  ...  In this paper, we propose a plug-and-play adversarial domain adaptation network (PnP-AdaNet) for adapting segmentation networks between different modalities of medical images, e.g., MRI and CT.  ...  ACKNOWLEDGMENT (Qi Dou and Cheng Ouyang contributed equally to this work.)  ... 
doi:10.1109/access.2019.2929258 fatcat:u4nuxyrzvzerfbrocgg44t6k5m

Unsupervised Bidirectional Cross-Modality Adaptation via Deeply Synergistic Image and Feature Alignment for Medical Image Segmentation [article]

Cheng Chen, Qi Dou, Hao Chen, Jing Qin, Pheng Ann Heng
2020 arXiv   pre-print
Unsupervised domain adaptation has increasingly gained interest in medical image computing, aiming to tackle the performance degradation of deep neural networks when being deployed to unseen data with  ...  The feature encoder is shared between both adaptive perspectives to leverage their mutual benefits via end-to-end learning.  ...  Accurate segmentation of multi-modality images can be achieved with deep learning techniques, but labeled data is required for each modality because it is difficult for deep models to generalize well across  ... 
arXiv:2002.02255v1 fatcat:hwqslvayxnh4tg37gmcso3o37u

Deep and Statistical Learning in Biomedical Imaging: State of the Art in 3D MRI Brain Tumor Segmentation [article]

K. Ruwani M. Fernando, Chris P. Tsokos
2021 arXiv   pre-print
Driven by the breakthroughs in computer vision, deep learning became the de facto standard in the domain of medical imaging.  ...  In this study, we critically review major statistical and deep learning models and their applications in brain imaging research with a focus on MRI-based brain tumor segmentation.  ...  While these methods generalize well to the unseen images, encoding prior information for a tumor is challenging.  ... 
arXiv:2103.05529v1 fatcat:iqu5ix5tgre6pnokdmoejywh74

Generalizable Cross-modality Medical Image Segmentation via Style Augmentation and Dual Normalization [article]

Ziqi Zhou, Lei Qi, Xin Yang, Dong Ni, Yinghuan Shi
2022 arXiv   pre-print
For medical image segmentation, imagine if a model was only trained using MR images in source domain, how about its performance to directly segment CT images in target domain?  ...  To be specific, given a single source domain, aiming to simulate the possible appearance change in unseen target domains, we first utilize a nonlinear transformation to augment source-similar and source-dissimilar  ...  [43] propose a deep-stacked transformation approach that employs a series of transformations to simulate domain shift for a specific medical imaging modality. Wang et al.  ... 
arXiv:2112.11177v3 fatcat:etqswsicn5fcfl57thg4kgkh6u
« Previous Showing results 1 — 15 out of 1,779 results