Category: IPMI 2021 oral

Authors: Sebastian G Popescu (Imperial College London)*; David Sharp (Imperial College London); James Cole (University College London); Konstantinos Kamnitsas (Imperial College London); Ben Glocker (Imperial College London)

Poster

Abstract: We propose a parameter efficient Bayesian layer for hierarchical convolutional Gaussian Processes that incorporates Gaussian Processes operating in Wasserstein-2 space to reliably propagate uncertainty. This directly replaces convolving Gaussian Processes with a distance-preserving affine operator on distributions. Our experiments on brain tissue-segmentation show that the resulting architecture approaches the performance of well-established deterministic segmentation algorithms (U-Net), which has never been achieved with previous hierarchical Gaussian Processes. Moreover, by applying the same segmentation model to out-of-distribution data (i.e., images with pathology such as brain tumors), we show that our uncertainty estimates result in out-of-distribution detection that outperforms the capabilities of previous Bayesian networks and reconstruction-based approaches that learn normative distributions.

Authors: Tom J van Sonsbeek (University of Amsterdam)*; Xiantong Zhen (University of Amsterdam); Marcel Worring (University of Amsterdam); Ling Shao (Inception Institute of Artificial Intelligence)

Poster

Abstract: Disease classification relying solely on imaging data attracts great interest in medical image analysis. Current models could be further improved, however, by also employing Electronic Health Records (EHRs), which contain rich information on patients and findings from clinicians. It is challenging to incorporate this information into disease classification due to the high reliance on clinician input in EHRs, limiting the possibility for automated diagnosis. In this paper, we propose variational knowledge distillation (VKD), which is a new probabilistic inference framework for disease classification based on X-rays that leverages knowledge from EHRs. Specifically, we introduce a conditional latent variable model, where we infer the latent representation of the X-ray image with the variational posterior conditioning on the associated EHR text. By doing so, the model acquires the ability to extract the visual features relevant to the disease during learning and can therefore perform more accurate classification for unseen patients at inference based solely on their X-ray scans. We demonstrate the effectiveness of our method on three public benchmark datasets with paired X-ray images and EHRs. The results show that the proposed variational knowledge distillation can consistently improve the performance of medical image classification and significantly surpasses current methods.

Authors: Sebastian Pölsterl (Ludwig-Maximilians Universität)*; Christian Wachinger (LMU Munich)

Poster

Abstract: Studying the relationship between neuroanatomy and cognitive decline due to Alzheimer’s has been a major research focus in the last decade. However, to infer cause-effect relationships rather than simple associations from observational data, we need to (i) express the causal relationships leading to cognitive decline in a graphical model, and (ii) ensure the causal effect of interest is identifiable from the collected data. We derive a causal graph from the current clinical knowledge on cause and effect in the Alzheimer’s disease continuum, and show that identifiability of the causal effect requires all confounders to be known and measured. However, in complex neuroimaging studies, we neither know all potential confounders nor do we have data on them. To alleviate this requirement, we leverage the dependencies among multiple causes by deriving a substitute confounder via a probabilistic latent factor model. In our theoretical analysis, we prove that using the substitute confounder enables identifiability of the causal effect of neuroanatomy on cognition. We quantitatively evaluate the effectiveness of our approach on semi-synthetic data, where we know the true causal effects, and illustrate its use on real data on the Alzheimer’s disease continuum, where it reveals important causes that otherwise would have been missed.

Authors: Andrew Hoopes (MGH)*; Malte Hoffmann (Harvard Medical School); Bruce Fischl (Massachusetts General Hospital / Harvard Medical School); John Guttag (MIT); Adrian V Dalca (MIT)

Poster

Abstract: We present HyperMorph, a learning-based strategy for deformable image registration that removes the need to tune important registration hyperparameters during training. Classical registration methods solve an optimization problem to find a set of spatial correspondences between two images, while learning-based methods leverage a training dataset to learn a function that generates these correspondences. The quality of the results for both types of techniques depends greatly on the choice of hyperparameters. Unfortunately, hyperparameter tuning is time-consuming and typically involves training many separate models with various hyperparameter values, potentially leading to suboptimal results. To address this inefficiency, we introduce amortized hyperparameter learning for image registration, a novel strategy to learn the effects of hyperparameters on deformation fields. The proposed framework learns a hypernetwork that takes in an input hyperparameter and modulates a registration network to produce the optimal deformation field for that hyperparameter value. In effect, this strategy trains a single, rich model that enables rapid, fine-grained discovery of hyperparameter values from a continuous interval at test-time. We demonstrate that this approach can be used to optimize multiple hyperparameters considerably faster than existing search strategies, leading to a reduced computational and human burden as well as increased flexibility. We also show several important benefits, including increased robustness to initialization and the ability to rapidly identify optimal hyperparameter values specific to a registration task, dataset, or even a single anatomical region, all without retraining the HyperMorph model. Our code is publicly available at http://voxelmorph.mit.edu.

Authors: Zixuan Liu (Stanford University); Ehsan Adeli (Stanford University); Kilian Pohl (Stanford University); Qingyu Zhao (Stanford University)*

Poster

Abstract: Interpretability is a critical factor in applying complex deep learning models to advance the understanding of brain disorders in neuroimaging studies. To interpret the decision process of a trained classifier, existing techniques typically rely on saliency maps to quantify the voxel-wise or feature-level importance for classification through partial derivatives. Despite providing some level of localization, these maps are not human-understandable from the neuroscience perspective as they often do not inform the specific type of morphological changes linked to the brain disorder. Inspired by the image-to-image translation scheme, we propose to train simulator networks to inject (or remove) patterns of the disease into a given MRI based on a warping operation, such that the classifier increases (or decreases) its confidence in labeling the simulated MRI as diseased. To increase the robustness of training, we propose to couple the two simulators into a unified model based on conditional convolution. We applied our approach to interpreting classifiers trained on a synthetic dataset and two neuroimaging datasets to visualize the effect of Alzheimer’s disease and alcohol dependence. Compared to the saliency maps generated by baseline approaches, our simulations and visualizations based on the Jacobian determinants of the warping field reveal meaningful and understandable patterns related to the diseases.

Authors: Kris Campbell (University of Utah)*; Haocheng Dai (University of Utah); Zhe Su (University of California, Los Angelese); Martin Bauer (Florida State University); Tom Fletcher (University of Virginia); Sarang Joshi (University of Utah, USA)

Poster

Abstract: The structural connectome is often represented by fiber bundles generated from various types of tractography. We propose a method of analyzing connectomes by representing them as a Riemannian metric, thereby viewing them as points in an infinite-dimensional manifold. After equipping this space with a natural metric structure, the Ebin metric, we apply object-oriented statistical analysis to define an atlas as the Fr\’echet mean of a population of Riemannian metrics. We demonstrate connectome registration and atlas formation using connectomes derived from diffusion tensors estimated from a subset of subjects from the Human Connectome Project.

Authors: Maëliss Jallais (Inria)*; Pedro Rodrigues (Inria); Alexandre Gramfort (Inria); Demian Wassermann (Inria)

Poster

Abstract: Effective characterisation of the brain grey matter cytoarchitecture with quantitative sensitivity to soma density and volume remains an unsolved challenge in diffusion MRI (dMRI). Solving the problem of relating the dMRI signal with cytoarchitectural characteristics calls for the definition of a mathematical model that describes brain tissue via a handful of physiologically-relevant parameters and an algorithm for inverting the model. To address this issue, we propose a new forward model, specifically a new system of equations, requiring six relatively sparse b-shells. These requirements are a drastic reduction of those used in current proposals to estimate grey matter cytoarchitecture. We then apply current tools from Bayesian analysis known as likelihood-free inference (LFI) to invert our proposed model. As opposed to other approaches from the literature, our LFI-based algorithm yields not only an estimation of the parameter vector that best describes a given observed data point, but also a full posterior distribution over the parameter space. This enables a richer description of the model inversion results providing indicators such as confidence intervals for the estimations, and better understanding of the parameter regions where the model may present indeterminacies. We approximate the posterior distribution using deep neural density estimators, known as normalizing flows, and fit them using a set of repeated simulations from the forward model. We validate our approach on simulations using dmipy and then apply the whole pipeline to the HCP MGH dataset.

Authors: Islem Mhiri (Université de Sousse); Ahmed Nebli (Higher Institute of Applied Science and Technologies (ISSAT), Universite de Sousse); Mohamed Ali Mahjoub (LATIS lab, National Engineering School of Sousse, ENISo, Sousse, Tunisia); Islem Rekik (Istanbul Technical University)*

Poster

Abstract: Brain graph synthesis marked a new era for predicting a target brain graph from a source one without incurring the high acquisition cost and processing time of neuroimaging data. However, existing multi-modal graph synthesis frameworks have several limitations. First, they mainly focus on generating graphs from the same domain (intra-modality), overlooking the rich multimodal representations of brain connectivity (inter-modality). Second, they can only handle isomorphic graph generation tasks, limiting their generalizability to synthesizing target graphs with a different node size and topological structure from those of the source one. More importantly, both target and source domains might have different distributions, which causes a domain fracture between them (i.e., distribution misalignment). To address such challenges, we propose an inter-modality aligner of non-isomorphic graphs (IMANGraphNet) framework to infer a target graph modality based on a given modality. Our three core contributions lie in (i) predicting a target graph (e.g., functional) from a source graph (e.g., morphological) based on a novel graph generative adversarial network (gGAN); (ii) using non-isomorphic graphs for both source and target domains with a different number of nodes, edges and structure; and (iii) enforcing the predicted target distribution to match that of the ground truth graphs using a graph autoencoder to relax the designed loss oprimization. Furthermore, to handle the unstable behavior of gGAN, we design a new Ground Truth-Preserving (GT-P) loss function to guide the generator in learning the topological structure of ground truth brain graphs more effectively. Our comprehensive experiments on predicting functional from morphological graphs demonstrate the outperformance of IMANGraphNet in comparison with its variants. This can be further leveraged for integrative and holistic brain mapping as well developing multimodal neurological diseases diagnostic frameworks.

Authors: Chun-Hao Yang (University of Florida)*; Baba Vemuri (University of Florida)

Poster

Abstract: Grassmann manifolds have been widely used to represent the geometry of feature spaces in a variety of problems in medical imaging and computer vision including but not limited to shape analysis, action recognition, subspace clustering and motion segmentation. For these problems, the features usually lie in a very high-dimensional Grassmann manifold and hence an appropriate dimensionality reduction technique is called for in order to curtail the computational burden. To this end, the Principal Geodesic Analysis (PGA), a nonlinear extension of the well known principal component analysis, is applicable as a general tool to many Riemannian manifolds. In this paper, we propose a novel framework for dimensionality reduction of data in Riemannian homogeneous spaces and then focus on the Grassman manifold which is an example of a homogeneous space. Our framework explicitly exploits the geometry of the homogeneous space yielding reduced dimensional nested sub-manifolds that need not be geodesic submanifolds and thus are more expressive. Specifically, we project points in a Grassmann manifold to an embedded lower dimensional Grassmann manifold. A salient feature of our method is that it leads to higher expressed variance compared to PGA which we demonstrate via synthetic and real data experiments.

Authors: Pierre-Louis A Antonsanti (GE Healthcare)*; Joan Glaunès (Université Paris 5); Thomas Benseghir (GE Healthcare); Vincent Jugnon (GE Healthcare); Irene Kaltenmark (MAP5 – Université de Paris)

Poster

Abstract:

In computer vision and medical imaging, the problem of matching structures finds numerous applications from automatic annotation to data reconstruction. The data however, while corresponding to the same anatomy, are often very different in topology or shape and might only partially match each other.
We introduce a new asymmetric data dissimilarity term for geometric shapes like curves, sets of curves or surfaces. This term is based on the Varifold shape representation and assess the embedding of a shape into another without relying on correspondences between points. It is designed as data attachment for the Large Deformation Diffeomorphic Metric Mapping (LDDMM), allowing to compute meaningful deformation of one shape onto a subset of the other. Registrations are illustrated on sets of synthetic 3D curves, real vascular trees and livers’ surfaces from two different modalities: Computed Tomography (CT) and Cone Beam Computed Tomography (CBCT). All experiments show that these data dissimilarity terms lead to coherent partial matching despite the topological differences.