Towards Segmentation and Spatial Alignment of the Human Embryonic Brain Using Deep Learning for Atlas-Based Registration

Bastiaansen, Wietske A.P.; Rousian, Melek; Steegers‐Theunissen, Régine P.M.; Niessen, Wiro J.; Koning, Anton H. J.; Klein, Stefan

doi:10.1007/978-3-030-50120-4_4

Cited by 8 publications

(11 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MI is used as the loss function. However, VoxelMorph consistently under-performed in our rigid registration task (similar is also observed by [62]) and we decided to not include the related results, for clarity.…”

Section: Voxelmorphmentioning

confidence: 56%

“…For sake of visual clarity we therefore omit the latter from the graph. As already commented, we also exclude results obtained by VoxelMorph [35] due to its demonstrated consistently poor performance, which we attribute to the constraint of rigid registration (as also observed by [62]). In Table 2, we summarise the aggregated performance, over all the considered displacements.…”

Section: Registration Performancementioning

confidence: 81%

See 1 more Smart Citation

Is image-to-image translation the panacea for multimodal image registration? A comparative study

Öfverstedt

Lindblad

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

Despite current advancement in the field of biomedical image processing, propelled by the deep learning revolution, multimodal image registration, due to its several challenges, is still often performed manually by specialists. The recent success of image-to-image (I2I) translation in computer vision applications and its growing use in biomedical areas provide a tempting possibility of transforming the multimodal registration problem into a, potentially easier, monomodal one. We conduct an empirical study of the applicability of modern I2I translation methods for the task of rigid registration of multimodal biomedical and medical 2D and 3D images. We compare the performance of four Generative Adversarial Network (GAN)-based I2I translation methods and one contrastive representation learning method, subsequently combined with two representative monomodal registration methods, to judge the effectiveness of modality translation for multimodal image registration. We evaluate these method combinations on four publicly available multimodal (2D and 3D) datasets and compare with the performance of registration achieved by several well-known approaches acting directly on multimodal image data. Our results suggest that, although I2I translation may be helpful when the modalities to register are clearly correlated, registration of modalities which express distinctly different properties of the sample are not well handled by the I2I translation approach. The evaluated representation learning method, which aims to find abstract image-like representations of the information shared between the modalities, manages better, and so does the Mutual Information maximisation approach, acting directly on the original multimodal images. We share our complete experimental setup as open-source (https://github.com/MIDA-group/MultiRegEval), including method implementations, evaluation code, and all datasets, for further reproducing and benchmarking.

show abstract

Section: Voxelmorphmentioning

confidence: 56%

Section: Registration Performancementioning

confidence: 81%

Is image-to-image translation the panacea for multimodal image registration? A comparative study

Öfverstedt

Lindblad

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Firstly, note that our framework consists of two networks, one dedicated to learning an affine transformation and the other to learning a nonrigid deformation. In previous work we showed that separating these tasks was needed due to the wide variety in positions and orientation of the embryo (Bastiaansen et al, 2020a). Therefore, the deformation φ consists of an affine transformation φ a and a nonrigid deformation φ d , such that:…”

Section: Learning To Estimate the Deformation φmentioning

confidence: 99%

“…( 10). When objects in the background are present, penalizing extreme zooming is beneficial, as was showed in Bastiaansen et al (2020a). Hence, L scaling is defined as:…”

Section: Loss Functionmentioning

confidence: 99%

Multi-Atlas Segmentation and Spatial Alignment of the Human Embryo in First Trimester 3D Ultrasound

Bastiaansen¹,

M.²,

Steegers‐Theunissen³

et al. 2022

Preprint

View full text Add to dashboard Cite

Segmentation and spatial alignment of ultrasound imaging data acquired in the in first trimester are crucial for monitoring early human embryonic growth and development throughout this crucial period of life. Current approaches are either manual or semi-automatic and are therefore very time-consuming and prone to errors. To automate these tasks, we propose a multi-atlas framework for automatic segmentation and spatial alignment of the embryo using deep learning with minimal supervision. Our framework learns to register the embryo to an atlas, which consists of the ultrasound images acquired at a range of gestational ages, segmented and spatially aligned to a predefined standard orientation. From this, we can derive the segmentation of the embryo and put the embryo in standard orientation. Ultrasound images acquired at 8+0 till 12+6 weeks gestational age were used and eight pregnancies were selected as atlas images. We evaluated different fusion strategies to incorporate multiple atlases: 1) training the framework using atlas images from a single pregnancy, 2) training the framework with data of all available atlases and 3) ensembling of the frameworks trained per pregnancy. To evaluate the performance, we calculated the Dice score over the test set. We found that training the framework using all available atlases outperformed ensembling and gave similar results compared to the best of all frameworks trained on a single subject. Furthermore, we found that selecting images from the four atlases closest in gestational age out of all available atlases, regardless of the individual quality, gave the best results with a median Dice score of 0.72. We conclude that our framework can accurately segment and spatially align the embryo in first trimester 3D ultrasound images and is robust for the variation in quality that existed in the available atlases. Our code is publicly available at: https://github.com/wapbastiaansen/multi-atlas-seg-reg.

show abstract

“…A popular combination strategy is to Email address: b.li@erasmusmc.nl (Bo Li) Abbreviations: MRI, Magnetic Resonance Imaging; DTI, Diffusion Tensor Imaging; FA, Fractional Anisotropy; MD, Mean Diffusivity; TE, Echo Time; TR, Repetition Time use the output of one task to optimize the other. Registration can be improved by using segmentation-level correspondences as input for deformation initialization (Dai and Khorram, 1999;Postelnicu et al, 2008) and optimization (De Groot et al, 2013b;Rohé et al, 2017;Hu et al, 2018;Balakrishnan et al, 2019;Bastiaansen et al, 2020;Zhu et al, 2020). Likewise, segmentation can benefit from registration by propagating anatomical information to subsequent frames, as has been shown in classical multi-atlas based segmentation methods (Fischl et al, 2002;Vakalopoulou et al, 2018) and in recent data-augmentation techniques which introduce labels to support unsupervised (Pathak et al, 2017) and weakly-supervised segmentation (Bortsova et al, 2019;Vlontzos and Mikolajczyk, 2018).…”

Section: Introductionmentioning

confidence: 99%

Longitudinal diffusion MRI analysis using Segis-Net: a single-step deep-learning framework for simultaneous segmentation and registration

Li,

Niessen,

Klein

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

This work presents a single-step deep-learning framework for longitudinal image analysis, coined Segis-Net. To optimally exploit information available in longitudinal data, this method concurrently learns a multi-class segmentation and nonlinear registration. Segmentation and registration are modeled using a convolutional neural network and optimized simultaneously for their mutual benefit. An objective function that optimizes spatial correspondence for the segmented structures across time-points is proposed. We applied Segis-Net to the analysis of white matter tracts from N=8045 longitudinal brain MRI datasets of 3249 elderly individuals. Segis-Net approach showed a significant increase in registration accuracy, spatio-temporal segmentation consistency, and reproducibility comparing with two multistage pipelines. This also led to a significant reduction in the sample-size that would be required to achieve the same statistical power in analyzing tract-specific measures. Thus, we expect that Segis-Net can serve as a new reliable tool to support longitudinal imaging studies to investigate macro-and microstructural brain changes over time.

show abstract

Towards Segmentation and Spatial Alignment of the Human Embryonic Brain Using Deep Learning for Atlas-Based Registration

Cited by 8 publications

References 16 publications

Is image-to-image translation the panacea for multimodal image registration? A comparative study

Is image-to-image translation the panacea for multimodal image registration? A comparative study

Multi-Atlas Segmentation and Spatial Alignment of the Human Embryo in First Trimester 3D Ultrasound

Longitudinal diffusion MRI analysis using Segis-Net: a single-step deep-learning framework for simultaneous segmentation and registration

Contact Info

Product

Resources

About