Yaël Balbastre

Sean I. Young, Yaël Balbastre, Adrian V. Dalca et al. · harvard

In recent years, learning-based image registration methods have gradually moved away from direct supervision with target warps to instead use self-supervision, with excellent results in several registration benchmarks. These approaches utilize a loss function that penalizes the intensity differences between the fixed and moving images, along with a suitable regularizer on the deformation. In this paper, we argue that the relative failure of supervised registration approaches can in part be blamed on the use of regular U-Nets, which are jointly tasked with feature extraction, feature matching, and estimation of deformation. We introduce one simple but crucial modification to the U-Net that disentangles feature extraction and matching from deformation prediction, allowing the U-Net to warp the features, across levels, as the deformation field is evolved. With this modification, direct supervision using target warps begins to outperform self-supervision approaches that require segmentations, presenting new directions for registration when images do not have segmentations. We hope that our findings in this preliminary workshop paper will re-ignite research interest in supervised image registration techniques. Our code is publicly available from https://github.com/balbasty/superwarp.

Mikael Brudfors, Yael Balbastre, John Ashburner et al. · harvard

Data used in image segmentation are not always defined on the same grid. This is particularly true for medical images, where the resolution, field-of-view and orientation can differ across channels and subjects. Images and labels are therefore commonly resampled onto the same grid, as a pre-processing step. However, the resampling operation introduces partial volume effects and blurring, thereby changing the effective resolution and reducing the contrast between structures. In this paper we propose a splat layer, which automatically handles resolution mismatches in the input data. This layer pushes each image onto a mean space where the forward pass is performed. As the splat operator is the adjoint to the resampling operator, the mean-space prediction can be pulled back to the native label space, where the loss function is computed. Thus, the need for explicit resolution adjustment using interpolation is removed. We show on two publicly available datasets, with simulated and real multi-modal magnetic resonance images, that this model improves segmentation results compared to resampling as a pre-processing step.

Matthew G. French, Gonzalo D. Maso Talou, Thiranja P. Babarenda Gamage et al. · harvard

In breast surgical planning, accurate registration of MR images across patient positions has the potential to improve the localisation of tumours during breast cancer treatment. While learning-based registration methods have recently become the state-of-the-art approach for most medical image registration tasks, these methods have yet to make inroads into breast image registration due to certain difficulties-the lack of rich texture information in breast MR images and the need for the deformations to be diffeomophic. In this work, we propose learning strategies for breast MR image registration that are amenable to diffeomorphic constraints, together with early experimental results from in-silico and in-vivo experiments. One key contribution of this work is a registration network which produces superior registration outcomes for breast images in addition to providing diffeomorphic guarantees.

Etienne Chollet, Yaël Balbastre, Chiara Mauri et al. · harvard

Microvascular anatomy is known to be involved in various neurological disorders. However, understanding these disorders is hindered by the lack of imaging modalities capable of capturing the comprehensive three-dimensional vascular network structure at microscopic resolution. With a lateral resolution of $<=$20 {\textmu}m and ability to reconstruct large tissue blocks up to tens of cubic centimeters, serial-section optical coherence tomography (sOCT) is well suited for this task. This method uses intrinsic optical properties to visualize the vessels and therefore does not possess a specific contrast, which complicates the extraction of accurate vascular models. The performance of traditional vessel segmentation methods is heavily degraded in the presence of substantial noise and imaging artifacts and is sensitive to domain shifts, while convolutional neural networks (CNNs) require extensive labeled data and are also sensitive the precise intensity characteristics of the data that they are trained on. Building on the emerging field of synthesis-based training, this study demonstrates a synthesis engine for neurovascular segmentation in sOCT images. Characterized by minimal priors and high variance sampling, our highly generalizable method tested on five distinct sOCT acquisitions eliminates the need for manual annotations while attaining human-level precision. Our approach comprises two phases: label synthesis and label-to-image transformation. We demonstrate the efficacy of the former by comparing it to several more realistic sets of training labels, and the latter by an ablation study of synthetic noise and artifact models.

Sean I. Young, Yaël Balbastre, Bruce Fischl et al. · harvard

In magnetic resonance imaging (MRI), slice-to-volume reconstruction (SVR) refers to computational reconstruction of an unknown 3D magnetic resonance volume from stacks of 2D slices corrupted by motion. While promising, current SVR methods require multiple slice stacks for accurate 3D reconstruction, leading to long scans and limiting their use in time-sensitive applications such as fetal fMRI. Here, we propose a SVR method that overcomes the shortcomings of previous work and produces state-of-the-art reconstructions in the presence of extreme inter-slice motion. Inspired by the recent success of single-view depth estimation methods, we formulate SVR as a single-stack motion estimation task and train a fully convolutional network to predict a motion stack for a given slice stack, producing a 3D reconstruction as a byproduct of the predicted motion. Extensive experiments on the SVR of adult and fetal brains demonstrate that our fully convolutional method is twice as accurate as previous SVR methods. Our code is available at github.com/seannz/svr.

Xiaoling Hu, Xiangrui Zeng, Oula Puonti et al. · harvard

Domain randomization through synthesis is a powerful strategy to train networks that are unbiased with respect to the domain of the input images. Randomization allows networks to see a virtually infinite range of intensities and artifacts during training, thereby minimizing overfitting to appearance and maximizing generalization to unseen data. Although powerful, this approach relies on the accurate tuning of a large set of hyperparameters that govern the probabilistic distribution of the synthesized images. Instead of manually tuning these parameters, we introduce Learn2Synth, a novel procedure in which synthesis parameters are learned using a small set of real labeled data. Unlike methods that impose constraints to align synthetic data with real data (e.g., contrastive or adversarial techniques), which risk misaligning the image and its label map, we tune an augmentation engine such that a segmentation network trained on synthetic data has optimal accuracy when applied to real data. This approach allows the training procedure to benefit from real labeled examples, without ever using these real examples to train the segmentation network, which avoids biasing the network towards the properties of the training set. Specifically, we develop parametric and nonparametric strategies to enhance synthetic images in a way that improves the performance of the segmentation network. We demonstrate the effectiveness of this learning strategy on synthetic and real-world brain scans. Code is available at: https://github.com/HuXiaoling/Learn2Synth.

Yaël Balbastre, Ali Aghaeifar, Nadège Corbin et al.

Purpose: Inter-scan motion is a substantial source of error in $R_1$ estimation, and can be expected to increase at 7T where $B_1$ fields are more inhomogeneous. The established correction scheme does not translate to 7T since it requires a body coil reference. Here we introduce two alternatives that outperform the established method. Since they compute relative sensitivities they do not require body coil images. Theory: The proposed methods use coil-combined magnitude images to obtain the relative coil sensitivities. The first method efficiently computes the relative sensitivities via a simple ratio; the second by fitting a more sophisticated generative model. Methods: $R_1$ maps were computed using the variable flip angle (VFA) approach. Multiple datasets were acquired at 3T and 7T, with and without motion between the acquisition of the VFA volumes. $R_1$ maps were constructed without correction, with the proposed corrections, and (at 3T) with the previously established correction scheme. Results: At 3T, the proposed methods outperform the baseline method. Inter-scan motion artefacts were also reduced at 7T. However, reproducibility only converged on that of the no motion condition if position-specific transmit field effects were also incorporated. Conclusion: The proposed methods simplify inter-scan motion correction of $R_1$ maps and are applicable at both 3T and 7T, where a body coil is typically not available. The open-source code for all methods is made publicly available.

Mikael Brudfors, Yaël Balbastre, John Ashburner et al.

While convolutional neural networks (CNNs) trained by back-propagation have seen unprecedented success at semantic segmentation tasks, they are known to struggle on out-of-distribution data. Markov random fields (MRFs) on the other hand, encode simpler distributions over labels that, although less flexible than UNets, are less prone to over-fitting. In this paper, we propose to fuse both strategies by computing the product of distributions of a UNet and an MRF. As this product is intractable, we solve for an approximate distribution using an iterative mean-field approach. The resulting MRF-UNet is trained jointly by back-propagation. Compared to other works using conditional random fields (CRFs), the MRF has no dependency on the imaging data, which should allow for less over-fitting. We show on 3D neuroimaging data that this novel network improves generalisation to out-of-distribution samples. Furthermore, it allows the overall number of parameters to be reduced while preserving high accuracy. These results suggest that a classic MRF smoothness prior can allow for less over-fitting when principally integrated into a CNN model. Our implementation is available at https://github.com/balbasty/nitorch.

Yael Balbastre, Mikael Brudfors, Michela Azzarito et al.

Quantitative MR imaging is increasingly favoured for its richer information content and standardised measures. However, computing quantitative parameter maps, such as those encoding longitudinal relaxation rate (R1), apparent transverse relaxation rate (R2*) or magnetisation-transfer saturation (MTsat), involves inverting a highly non-linear function. Many methods for deriving parameter maps assume perfect measurements and do not consider how noise is propagated through the estimation procedure, resulting in needlessly noisy maps. Instead, we propose a probabilistic generative (forward) model of the entire dataset, which is formulated and inverted to jointly recover (log) parameter maps with a well-defined probabilistic interpretation (e.g., maximum likelihood or maximum a posteriori). The second order optimisation we propose for model fitting achieves rapid and stable convergence thanks to a novel approximate Hessian. We demonstrate the utility of our flexible framework in the context of recovering more accurate maps from data acquired using the popular multi-parameter mapping protocol. We also show how to incorporate a joint total variation prior to further decrease the noise in the maps, noting that the probabilistic formulation allows the uncertainty on the recovered parameter maps to be estimated. Our implementation uses a PyTorch backend and benefits from GPU acceleration. It is available at https://github.com/balbasty/nitorch.

Juan Eugenio Iglesias, Benjamin Billot, Yael Balbastre et al.

Most existing algorithms for automatic 3D morphometry of human brain MRI scans are designed for data with near-isotropic voxels at approximately 1 mm resolution, and frequently have contrast constraints as well - typically requiring T1 scans (e.g., MP-RAGE). This limitation prevents the analysis of millions of MRI scans acquired with large inter-slice spacing ("thick slice") in clinical settings every year. The inability to quantitatively analyze these scans hinders the adoption of quantitative neuroimaging in healthcare, and precludes research studies that could attain huge sample sizes and hence greatly improve our understanding of the human brain. Recent advances in CNNs are producing outstanding results in super-resolution and contrast synthesis of MRI. However, these approaches are very sensitive to the contrast, resolution and orientation of the input images, and thus do not generalize to diverse clinical acquisition protocols - even within sites. Here we present SynthSR, a method to train a CNN that receives one or more thick-slice scans with different contrast, resolution and orientation, and produces an isotropic scan of canonical contrast (typically a 1 mm MP-RAGE). The presented method does not require any preprocessing, e.g., skull stripping or bias field correction. Crucially, SynthSR trains on synthetic input images generated from 3D segmentations, and can thus be used to train CNNs for any combination of contrasts, resolutions and orientations without high-resolution training data. We test the images generated with SynthSR in an array of common downstream analyses, and show that they can be reliably used for subcortical segmentation and volumetry, image registration (e.g., for tensor-based morphometry), and, if some image quality requirements are met, even cortical thickness morphometry. The source code is publicly available at github.com/BBillot/SynthSR.

Mikael Brudfors, Yael Balbastre, Parashkev Nachev et al.

We present a tool for resolution recovery in multimodal clinical magnetic resonance imaging (MRI). Such images exhibit great variability, both biological and instrumental. This variability makes automated processing with neuroimaging analysis software very challenging. This leaves intelligence extractable only from large-scale analyses of clinical data untapped, and impedes the introduction of automated predictive systems in clinical care. The tool presented in this paper enables such processing, via inference in a generative model of thick-sliced, multi-contrast MR scans. All model parameters are estimated from the observed data, without the need for manual tuning. The model-driven nature of the approach means that no type of training is needed for applicability to the diversity of MR contrasts present in a clinical context. We show on simulated data that the proposed approach outperforms conventional model-based techniques, and on a large hospital dataset of multimodal MRIs that the tool can successfully super-resolve very thick-sliced images. The implementation is available from https://github.com/brudfors/spm_superres.

Mikael Brudfors, Yael Balbastre, Parashkev Nachev et al.

This paper presents a generative model for super-resolution in routine clinical magnetic resonance images (MRI), of arbitrary orientation and contrast. The model recasts the recovery of high resolution images as an inverse problem, in which a forward model simulates the slice-select profile of the MR scanner. The paper introduces a prior based on multi-channel total variation for MRI super-resolution. Bias-variance trade-off is handled by estimating hyper-parameters from the low resolution input scans. The model was validated on a large database of brain images. The validation showed that the model can improve brain segmentation, that it can recover anatomical information between images of different MR contrasts, and that it generalises well to the large variability present in MR images of different subjects. The implementation is freely available at https://github.com/brudfors/spm_superres

Etienne Chollet, Yael Balbastre, Caroline Magnain et al. · harvard

Serial sectioning Optical Coherence Tomography (sOCT) is a high-throughput, label free microscopic imaging technique that is becoming increasingly popular to study post-mortem neurovasculature. Quantitative analysis of the vasculature requires highly accurate segmentation; however, sOCT has low signal-to-noise-ratio and displays a wide range of contrasts and artifacts that depend on acquisition parameters. Furthermore, labeled data is scarce and extremely time consuming to generate. Here, we leverage synthetic datasets of vessels to train a deep learning segmentation model. We construct the vessels with semi-realistic splines that simulate the vascular geometry and compare our model with realistic vascular labels generated by constrained constructive optimization. Both approaches yield similar Dice scores, although with very different false positive and false negative rates. This method addresses the complexity inherent in OCT images and paves the way for more accurate and efficient analysis of neurovascular structures.

Alessa Hering, Lasse Hansen, Tony C. W. Mok et al.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https://learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Yu Yan, Yael Balbastre, Mikael Brudfors et al.

Segmentation of brain magnetic resonance images (MRI) into anatomical regions is a useful task in neuroimaging. Manual annotation is time consuming and expensive, so having a fully automated and general purpose brain segmentation algorithm is highly desirable. To this end, we propose a patched-based label propagation approach based on a generative model with latent variables. Once trained, our Factorisation-based Image Labelling (FIL) model is able to label target images with a variety of image contrasts. We compare the effectiveness of our proposed model against the state-of-the-art using data from the MICCAI 2012 Grand Challenge and Workshop on Multi-Atlas Labeling. As our approach is intended to be general purpose, we also assess how well it can handle domain shift by labelling images of the same subjects acquired with different MR contrasts.

Mikael Brudfors, John Ashburner, Parashkev Nachev et al.

Automatically generating one medical imaging modality from another is known as medical image translation, and has numerous interesting applications. This paper presents an interpretable generative modelling approach to medical image translation. By allowing a common model for group-wise normalisation and segmentation of brain scans to handle missing data, the model allows for predicting entirely missing modalities from one, or a few, MR contrasts. Furthermore, the model can be trained on a fairly small number of subjects. The proposed model is validated on three clinically relevant scenarios. Results appear promising and show that a principled, probabilistic model of the relationship between multi-channel signal intensities can be used to infer missing modalities -- both MR contrasts and CT images.

Mikael Brudfors, Yaël Balbastre, John Ashburner

Although convolutional neural networks (CNNs) currently dominate competitions on image segmentation, for neuroimaging analysis tasks, more classical generative approaches based on mixture models are still used in practice to parcellate brains. To bridge the gap between the two, in this paper we propose a marriage between a probabilistic generative model, which has been shown to be robust to variability among magnetic resonance (MR) images acquired via different imaging protocols, and a CNN. The link is in the prior distribution over the unknown tissue classes, which are classically modelled using a Markov random field. In this work we model the interactions among neighbouring pixels by a type of recurrent CNN, which can encode more complex spatial interactions. We validate our proposed model on publicly available MR data, from different centres, and show that it generalises across imaging protocols. This result demonstrates a successful and principled inclusion of a CNN in a generative model, which in turn could be adapted by any probabilistic generative approach for image segmentation.

John Ashburner, Mikael Brudfors, Kevin Bronik et al.

This paper presents a framework for automatically learning shape and appearance models for medical (and certain other) images. It is based on the idea that having a more accurate shape and appearance model leads to more accurate image registration, which in turn leads to a more accurate shape and appearance model. This leads naturally to an iterative scheme, which is based on a probabilistic generative model that is fit using Gauss-Newton updates within an EM-like framework. It was developed with the aim of enabling distributed privacy-preserving analysis of brain image data, such that shared information (shape and appearance basis functions) may be passed across sites, whereas latent variables that encode individual images remain secure within each site. These latent variables are proposed as features for privacy-preserving data mining applications. The approach is demonstrated qualitatively on the KDEF dataset of 2D face images, showing that it can align images that traditionally require shape and appearance models trained using manually annotated data (manually defined landmarks etc.). It is applied to MNIST dataset of handwritten digits to show its potential for machine learning applications, particularly when training data is limited. The model is able to handle ``missing data'', which allows it to be cross-validated according to how well it can predict left-out voxels. The suitability of the derived features for classifying individuals into patient groups was assessed by applying it to a dataset of over 1,900 segmented T1-weighted MR images, which included images from the COBRE and ABIDE datasets.