We propose an efficient non-parametric diffeomorphic image registration algorithm based on Thirion's demons algorithm. In the first part of this paper, we show that Thirion's demons algorithm can be ...seen as an optimization procedure on the entire space of displacement fields. We provide strong theoretical roots to the different variants of Thirion's demons algorithm. This analysis predicts a theoretical advantage for the symmetric forces variant of the demons algorithm. We show on controlled experiments that this advantage is confirmed in practice and yields a faster convergence. In the second part of this paper, we adapt the optimization procedure underlying the demons algorithm to a space of diffeomorphic transformations. In contrast to many diffeomorphic registration algorithms, our solution is computationally efficient since in practice it only replaces an addition of displacement fields by a few compositions. Our experiments show that in addition to being diffeomorphic, our algorithm provides results that are similar to the ones from the demons algorithm but with transformations that are much smoother and closer to the gold standard, available in controlled experiments, in terms of Jacobians.
Automatic segmentation of brain tumors from medical images is important for clinical assessment and treatment planning of brain tumors. Recent years have seen an increasing use of convolutional ...neural networks (CNNs) for this task, but most of them use either 2D networks with relatively low memory requirement while ignoring 3D context, or 3D networks exploiting 3D features while with large memory consumption. In addition, existing methods rarely provide uncertainty information associated with the segmentation result. We propose a cascade of CNNs to segment brain tumors with hierarchical subregions from multi-modal Magnetic Resonance images (MRI), and introduce a 2.5D network that is a trade-off between memory consumption, model complexity and receptive field. In addition, we employ test-time augmentation to achieve improved segmentation accuracy, which also provides voxel-wise and structure-wise uncertainty information of the segmentation result. Experiments with BraTS 2017 dataset showed that our cascaded framework with 2.5D CNNs was one of the top performing methods (second-rank) for the BraTS challenge. We also validated our method with BraTS 2018 dataset and found that test-time augmentation improves brain tumor segmentation accuracy and that the resulting uncertainty information can indicate potential mis-segmentations and help to improve segmentation accuracy.
Accurate medical image segmentation is essential for diagnosis, surgical planning and many other applications. Convolutional Neural Networks (CNNs) have become the state-of-the-art automatic ...segmentation methods. However, fully automatic results may still need to be refined to become accurate and robust enough for clinical use. We propose a deep learning-based interactive segmentation method to improve the results obtained by an automatic CNN and to reduce user interactions during refinement for higher accuracy. We use one CNN to obtain an initial automatic segmentation, on which user interactions are added to indicate mis-segmentations. Another CNN takes as input the user interactions with the initial segmentation and gives a refined result. We propose to combine user interactions with CNNs through geodesic distance transforms, and propose a resolution-preserving network that gives a better dense prediction. In addition, we integrate user interactions as hard constraints into a back-propagatable Conditional Random Field. We validated the proposed framework in the context of 2D placenta segmentation from fetal MRI and 3D brain tumor segmentation from FLAIR images. Experimental results show our method achieves a large improvement from automatic CNNs, and obtains comparable and even higher accuracy with fewer user interventions and less time compared with traditional interactive methods.
High-resolution volume reconstruction from multiple motion-corrupted stacks of 2D slices plays an increasing role for fetal brain Magnetic Resonance Imaging (MRI) studies. Currently existing ...reconstruction methods are time-consuming and often require user interactions to localize and extract the brain from several stacks of 2D slices. We propose a fully automatic framework for fetal brain reconstruction that consists of four stages: 1) fetal brain localization based on a coarse segmentation by a Convolutional Neural Network (CNN), 2) fine segmentation by another CNN trained with a multi-scale loss function, 3) novel, single-parameter outlier-robust super-resolution reconstruction, and 4) fast and automatic high-resolution visualization in standard anatomical space suitable for pathological brains. We validated our framework with images from fetuses with normal brains and with variable degrees of ventriculomegaly associated with open spina bifida, a congenital malformation affecting also the brain. Experiments show that each step of our proposed pipeline outperforms state-of-the-art methods in both segmentation and reconstruction comparisons including expert-reader quality assessments. The reconstruction results of our proposed method compare favorably with those obtained by manual, labor-intensive brain segmentation, which unlocks the potential use of automatic fetal brain reconstruction studies in clinical practice.
•A method to infer voxel-level correspondence from higher-level anatomical labels.•Efficient and fully-automated registration for MR and ultrasound prostate images.•Validation experiments with 108 ...pairs of labelled interventional patient images.•Open-source implementation.
One of the fundamental challenges in supervised learning for multimodal image registration is the lack of ground-truth for voxel-level spatial correspondence. This work describes a method to infer voxel-level transformation from higher-level correspondence information contained in anatomical labels. We argue that such labels are more reliable and practical to obtain for reference sets of image pairs than voxel-level correspondence. Typical anatomical labels of interest may include solid organs, vessels, ducts, structure boundaries and other subject-specific ad hoc landmarks. The proposed end-to-end convolutional neural network approach aims to predict displacement fields to align multiple labelled corresponding structures for individual image pairs during the training, while only unlabelled image pairs are used as the network input for inference. We highlight the versatility of the proposed strategy, for training, utilising diverse types of anatomical labels, which need not to be identifiable over all training image pairs. At inference, the resulting 3D deformable image registration algorithm runs in real-time and is fully-automated without requiring any anatomical labels or initialisation. Several network architecture variants are compared for registering T2-weighted magnetic resonance images and 3D transrectal ultrasound images from prostate cancer patients. A median target registration error of 3.6 mm on landmark centroids and a median Dice of 0.87 on prostate glands are achieved from cross-validation experiments, in which 108 pairs of multimodal images from 76 patients were tested with high-quality anatomical labels.
Display omitted
Symmetric log-domain diffeomorphic Registration: a demons-based approach Vercauteren, Tom; Pennec, Xavier; Perchant, Aymeric ...
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention,
2008, Letnik:
11, Številka:
Pt 1
Journal Article, Conference Proceeding
Recenzirano
Odprti dostop
Modern morphometric studies use non-linear image registration to compare anatomies and perform group analysis. Recently, log-Euclidean approaches have contributed to promote the use of such ...computational anatomy tools by permitting simple computations of statistics on a rather large class of invertible spatial transformations. In this work, we propose a non-linear registration algorithm perfectly fit for log-Euclidean statistics on diffeomorphisms. Our algorithm works completely in the log-domain, i.e. it uses a stationary velocity field. This implies that we guarantee the invertibility of the deformation and have access to the true inverse transformation. This also means that our output can be directly used for log-Euclidean statistics without relying on the heavy computation of the log of the spatial transformation. As it is often desirable, our algorithm is symmetric with respect to the order of the input images. Furthermore, we use an alternate optimization approach related to Thirion's demons algorithm to provide a fast non-linear registration algorithm. First results show that our algorithm outperforms both the demons algorithm and the recently proposed diffeomorphic demons algorithm in terms of accuracy of the transformation while remaining computationally efficient.
Photoacoustic imaging (PAI) is an emerging biomedical imaging modality that is based on optical absorption contrast, capable of revealing distinct spectroscopic signatures of tissue at high spatial ...resolution and large imaging depths. However, clinical applications of conventional non-invasive PAI systems have been restricted to examinations of tissues at depths less than a few cm due to strong light attenuation. Minimally invasive photoacoustic imaging (miPAI) has greatly extended the landscape of PAI by delivering excitation light within tissue through miniature fibre-optic probes. In the past decade, various miPAI systems have been developed with demonstrated applicability in several clinical fields. In this article, we present an overview of the current status of miPAI and our thoughts on future perspectives.
Objective
Clinical outcomes in multiple sclerosis (MS) are highly variable. We aim to determine the long‐term clinical outcomes in MS, and to identify early prognostic features of these outcomes.
...Methods
One hundred thirty‐two people presenting with a clinically isolated syndrome were prospectively recruited between 1984 and 1987, and followed up clinically and radiologically 1, 5, 10, 14, 20, and now 30 years later. All available notes and magnetic resonance imaging scans were reviewed, and MS was defined according to the 2010 McDonald criteria.
Results
Clinical outcome data were obtained in 120 participants at 30 years. Eighty were known to have developed MS by 30 years. Expanded Disability Status Scale (EDSS) scores were available in 107 participants, of whom 77 had MS; 32 (42%) remained fully ambulatory (EDSS scores ≤3.5), all of whom had relapsing–remitting MS (RRMS), 3 (4%) had RRMS and EDSS scores >3.5, 26 (34%) had secondary progressive MS (all had EDSS scores >3.5), and MS contributed to death in 16 (20%). Of those with MS, 11 received disease‐modifying therapy. The strongest early predictors (within 5 years of presentation) of secondary progressive MS at 30 years were presence of baseline infratentorial lesions and deep white matter lesions at 1 year.
Interpretation
Thirty years after onset, in a largely untreated cohort, there was a divergence of MS outcomes; some people accrued substantial disability early on, whereas others ran a more favorable long‐term course. These outcomes could, in part, be predicted by radiological findings from within 1 year of first presentation. ANN NEUROL 2020;87:63–74
All fields of neuroscience that employ brain imaging need to communicate their results with reference to anatomical regions. In particular, comparative morphometry and group analysis of functional ...and physiological data require coregistration of brains to establish correspondences across brain structures. It is well established that linear registration of one brain to another is inadequate for aligning brain structures, so numerous algorithms have emerged to nonlinearly register brains to one another. This study is the largest evaluation of nonlinear deformation algorithms applied to brain image registration ever conducted. Fourteen algorithms from laboratories around the world are evaluated using 8 different error measures. More than 45,000 registrations between 80 manually labeled brains were performed by algorithms including: AIR, ANIMAL, ART, Diffeomorphic Demons, FNIRT, IRTK, JRD-fluid, ROMEO, SICLE, SyN, and four different SPM5 algorithms (“SPM2-type” and regular Normalization, Unified Segmentation, and the DARTEL Toolbox). All of these registrations were preceded by linear registration between the same image pairs using FLIRT. One of the most significant findings of this study is that the relative performances of the registration methods under comparison appear to be little affected by the choice of subject population, labeling protocol, and type of overlap measure. This is important because it suggests that the findings are generalizable to new subject populations that are labeled or evaluated using different labeling protocols. Furthermore, we ranked the 14 methods according to three completely independent analyses (permutation tests, one-way ANOVA tests, and indifference-zone ranking) and derived three almost identical top rankings of the methods. ART, SyN, IRTK, and SPM's DARTEL Toolbox gave the best results according to overlap and distance measures, with ART and SyN delivering the most consistently high accuracy across subjects and label sets. Updates will be published on the http://www.mindboggle.info/papers/ website.