Search

Recently, deep reinforcement learning (RL) has been proposed to learn the tractography procedure and train agents to reconstruct the structure of the white matter without manually curated reference streamlines. While the performances reported were competitive, the proposed framework is complex, and little is still known about the role and impact of its multiple parts. In this work, we thoroughly explore the different components of the proposed framework, such as the choice of the RL algorithm, seeding strategy, the input signal and reward function, and shed light on their impact. Approximately 7,400 models were trained for this work, totalling nearly 41,000 h of GPU time. Our goal is to guide researchers eager to explore the possibilities of deep RL for tractography by exposing what works and what does not work with the category of approach. As such, we ultimately propose a series of recommendations concerning the choice of RL algorithm, the input to the agents, the reward function and more to help future work using reinforcement learning for tractography. We also release the open source codebase, trained models, and datasets for users and researchers wanting to explore reinforcement learning for tractography., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier B.V. All rights reserved.)

White matter bundle segmentation is a cornerstone of modern tractography to study the brain's structural connectivity in domains such as neurological disorders, neurosurgery, and aging. In this study, we present FIESTA (FIbEr Segmentation in Tractography using Autoencoders), a reliable and robust, fully automated, and easily semi-automatically calibrated pipeline based on deep autoencoders that can dissect and fully populate white matter bundles. This pipeline is built upon previous works that demonstrated how autoencoders can be used successfully for streamline filtering, bundle segmentation, and streamline generation in tractography. Our proposed method improves bundle segmentation coverage by recovering hard-to-track bundles with generative sampling through the latent space seeding of the subject bundle and the atlas bundle. A latent space of streamlines is learned using autoencoder-based modeling combined with contrastive learning. Using an atlas of bundles in standard space (MNI), our proposed method segments new tractograms using the autoencoder latent distance between each tractogram streamline and its closest neighbor bundle in the atlas of bundles. Intra-subject bundle reliability is improved by recovering hard-to-track streamlines, using the autoencoder to generate new streamlines that increase the spatial coverage of each bundle while remaining anatomically correct. Results show that our method is more reliable than state-of-the-art automated virtual dissection methods such as RecoBundles, RecoBundlesX, TractSeg, White Matter Analysis and XTRACT. Our framework allows for the transition from one anatomical bundle definition to another with marginal calibration efforts. Overall, these results show that our framework improves the practicality and usability of current state-of-the-art bundle segmentation framework., Competing Interests: Declaration of competing interest 1. Maxime Descoteaux and Pierre-Marc Jodoin report membership and employment with Imeka Solutions inc. 2. Patent #17/337,413 is pending to Imeka Solutions inc. with inventors Jon Haitz Legarreta, Maxime Descoteaux and Pierre-Marc Jodoin. 3. Muhamed Barakovic is an employee of Hays plc and a consultant for F. Hoffmann-La Roche Ltd. 4. Stefano Magon is an employee and shareholder of F. Hoffmann-La Roche Ltd., (Copyright © 2023 The Authors. Published by Elsevier Inc. All rights reserved.)

Current tractography methods use the local orientation information to propagate streamlines from seed locations. Many such seeds provide streamlines that stop prematurely or fail to map the true white matter pathways because some bundles are "harder-to-track" than others. This results in tractography reconstructions with poor white and gray matter spatial coverage. In this work, we propose a generative, autoencoder-based method, named GESTA (Generative Sampling in Bundle Tractography using Autoencoders), that produces streamlines achieving better spatial coverage. Compared to other deep learning methods, our autoencoder-based framework uses a single model to generate streamlines in a bundle-wise fashion, and does not require to propagate local orientations. GESTA produces new and complete streamlines for any given white matter bundle, including hard-to-track bundles. Applied on top of a given tractogram, GESTA is shown to be effective in improving the white matter volume coverage in poorly populated bundles, both on synthetic and human brain in vivo data. Our streamline evaluation framework ensures that the streamlines produced by GESTA are anatomically plausible and fit well to the local diffusion signal. The streamline evaluation criteria assess anatomy (white matter coverage), local orientation alignment (direction), and geometry features of streamlines, and optionally, gray matter connectivity. The GESTA framework offers considerable gains in bundle overlap using a reduced set of seeding streamlines with a 1.5x improvement for the "Fiber Cup", and 6x for the ISMRM 2015 Tractography Challenge datasets. Similarly, it provides a 4x white matter volume increase on the BIL&GIN callosal homotopic dataset, and successfully populates bundles on the multi-subject, multi-site, whole-brain in vivo TractoInferno dataset. GESTA is thus a novel deep generative bundle tractography method that can be used to improve the tractography reconstruction of the white matter., Competing Interests: Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jon Haitz Legarreta has patent #17/337,413 pending to Imeka Solutions inc. Pierre-Marc Jodoin reports a relationship with Imeka Solutions inc. that includes: board membership and employment. Pierre-Marc Jodoin has patent #17/337,413 pending to Imeka Solutions inc. Maxime Descoteaux reports a relationship with Imeka Solutions inc. that includes: board membership and employment. Maxime Descoteaux has patent #17/337,413 pending to Imeka Solutions inc., (Copyright © 2023 Elsevier B.V. All rights reserved.)

Generative adversarial networks (GANs) have become increasingly powerful, generating mind-blowing photorealistic images that mimic the content of datasets they have been trained to replicate. One recurrent theme in medical imaging, is whether GANs can also be as effective at generating workable medical data, as they are for generating realistic RGB images. In this paper, we perform a multi-GAN and multi-application study, to gauge the benefits of GANs in medical imaging. We tested various GAN architectures, from basic DCGAN to more sophisticated style-based GANs, on three medical imaging modalities and organs, namely: cardiac cine-MRI, liver CT, and RGB retina images. GANs were trained on well-known and widely utilized datasets, from which their FID scores were computed, to measure the visual acuity of their generated images. We further tested their usefulness by measuring the segmentation accuracy of a U-Net trained on these generated images and the original data. The results reveal that GANs are far from being equal, as some are ill-suited for medical imaging applications, while others performed much better. The top-performing GANs are capable of generating realistic-looking medical images by FID standards, that can fool trained experts in a visual Turing test and comply to some metrics. However, segmentation results suggest that no GAN is capable of reproducing the full richness of medical datasets.

TractoInferno is the world's largest open-source multi-site tractography database, including both research- and clinical-like human acquisitions, aimed specifically at machine learning tractography approaches and related ML algorithms. It provides 284 samples acquired from 3 T scanners across 6 different sites. Available data includes T1-weighted images, single-shell diffusion MRI (dMRI) acquisitions, spherical harmonics fitted to the dMRI signal, fiber ODFs, and reference streamlines for 30 delineated bundles generated using 4 tractography algorithms, as well as masks needed to run tractography algorithms. Manual quality control was additionally performed at multiple steps of the pipeline. We showcase TractoInferno by benchmarking the learn2track algorithm and 5 variations of the same recurrent neural network architecture. Creating the TractoInferno database required approximately 20,000 CPU-hours of processing power, 200 man-hours of manual QC, 3,000 GPU-hours of training baseline models, and 4 Tb of storage, to produce a final database of 350 Gb. By providing a standardized training dataset and evaluation protocol, TractoInferno is an excellent tool to address common issues in machine learning tractography., (© 2022. The Author(s).)

Convolutional neural networks (CNN) have demonstrated their ability to segment 2D cardiac ultrasound images. However, despite recent successes according to which the intra-observer variability on end-diastole and end-systole images has been reached, CNNs still struggle to leverage temporal information to provide accurate and temporally consistent segmentation maps across the whole cycle. Such consistency is required to accurately describe the cardiac function, a necessary step in diagnosing many cardiovascular diseases. In this paper, we propose a framework to learn the 2D+time apical long-axis cardiac shape such that the segmented sequences can benefit from temporal and anatomical consistency constraints. Our method is a post-processing that takes as input segmented echocardiographic sequences produced by any state-of-the-art method and processes it in two steps to (i) identify spatio-temporal inconsistencies according to the overall dynamics of the cardiac sequence and (ii) correct the inconsistencies. The identification and correction of cardiac inconsistencies relies on a constrained autoencoder trained to learn a physiologically interpretable embedding of cardiac shapes, where we can both detect and fix anomalies. We tested our framework on 98 full-cycle sequences from the CAMUS dataset, which are available alongside this paper. Our temporal regularization method not only improves the accuracy of the segmentation across the whole sequences, but also enforces temporal and anatomical consistency.

Modern tractography algorithms such as anatomically-constrained tractography (ACT) are based on segmentation maps of white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). These maps are generally estimated from a T1-weighted (T1w) image and then registered in diffusion weighted images (DWI) space. Registration of T1w to diffusion space and partial volume estimation are challenging and rarely voxel-perfect. Diffusion-based segmentation would, thus, potentially allow not to have higher quality anatomical priors injected in the tractography process. On the other hand, even if FA-based tractography is possible without T1 registration, the literature shows that this technique suffers from multiple issues such as holes in the tracking mask and a high proportion of generated broken and anatomically implausible streamlines. Therefore, there is an important need for a tissue segmentation algorithm that works directly in the native diffusion space. We propose DORIS , a DWI-based deep learning segmentation algorithm. DORIS outputs 10 different tissue classes including WM, GM, CSF, ventricles, and 6 other subcortical structures (putamen, pallidum, hippocampus, caudate, amygdala, and thalamus). DORIS was trained and validated on a wide range of subjects, including 1,000 individuals from 22 to 90 years old from clinical and research DWI acquisitions, from 5 public databases. In the absence of a "true" ground truth in diffusion space, DORIS used a silver standard strategy from Freesurfer output registered onto the DWI. This strategy is extensively evaluated and discussed in the current study. Segmentation maps provided by DORIS are quantitatively compared to Freesurfer and FSL-fast and the impacts on tractography are evaluated. Overall, we show that DORIS is fast, accurate, and reproducible and that DORIS-based tractograms produce bundles with a longer mean length and fewer anatomically implausible streamlines., Competing Interests: Authors GT, MDu, CZ, P-MJ, and MDe are employed by the company Imeka Solutions Inc. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Theaud, Edde, Dumont, Zotti, Zucchelli, Deslauriers-Gauthier, Deriche, Jodoin and Descoteaux.)

The physical and clinical constraints surrounding diffusion-weighted imaging (DWI) often limit the spatial resolution of the produced images to voxels up to eight times larger than those of T1w images. The detailed information contained in accessible high-resolution T1w images could help in the synthesis of diffusion images with a greater level of detail. However, the non-Euclidean nature of diffusion imaging hinders current deep generative models from synthesizing physically plausible images. In this work, we propose the first Riemannian network architecture for the direct generation of diffusion tensors (DT) and diffusion orientation distribution functions (dODFs) from high-resolution T1w images. Our integration of the log-Euclidean Metric into a learning objective guarantees, unlike standard Euclidean networks, the mathematically-valid synthesis of diffusion. Furthermore, our approach improves the fractional anisotropy mean squared error (FA MSE) between the synthesized diffusion and the ground-truth by more than 23% and the cosine similarity between principal directions by almost 5% when compared to our baselines. We validate our generated diffusion by comparing the resulting tractograms to our expected real data. We observe similar fiber bundles with streamlines having <3% difference in length, <1% difference in volume, and a visually close shape. While our method is able to generate diffusion images from structural inputs in a high-resolution space within 15 s, we acknowledge and discuss the limits of diffusion inference solely relying on T1w images. Our results nonetheless suggest a relationship between the high-level geometry of the brain and its overall white matter architecture that remains to be explored., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Anctil-Robitaille, Théberge, Jodoin, Descoteaux, Desrosiers and Lombaert.)

Multiparametric magnetic resonance imaging (mp-MRI) has shown excellent results in the detection of prostate cancer (PCa). However, characterizing prostate lesions aggressiveness in mp-MRI sequences is impossible in clinical practice, and biopsy remains the reference to determine the Gleason score (GS). In this work, we propose a novel end-to-end multi-class network that jointly segments the prostate gland and cancer lesions with GS group grading. After encoding the information on a latent space, the network is separated in two branches: 1) the first branch performs prostate segmentation 2) the second branch uses this zonal prior as an attention gate for the detection and grading of prostate lesions. The model was trained and validated with a 5-fold cross-validation on a heterogeneous series of 219 MRI exams acquired on three different scanners prior prostatectomy. In the free-response receiver operating characteristics (FROC) analysis for clinically significant lesions (defined as GS >6) detection, our model achieves 69.0%±14.5% sensitivity at 2.9 false positive per patient on the whole prostate and  70.8%±14.4% sensitivity at 1.5 false positive when considering the peripheral zone (PZ) only. Regarding the automatic GS group grading, Cohen's quadratic weighted kappa coefficient (κ) is 0.418±0.138, which is the best reported lesion-wise kappa for GS segmentation to our knowledge. The model has encouraging generalization capacities with κ=0.120±0.092 on the PROSTATEx-2 public dataset and achieves state-of-the-art performance for the segmentation of the whole prostate gland with a Dice of 0.875±0.013. Finally, we show that ProstAttention-Net improves performance in comparison to reference segmentation models, including U-Net, DeepLabv3+ and E-Net. The proposed attention mechanism is also shown to outperform Attention U-Net., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2021. Published by Elsevier B.V.)

Generative adversarial networks (GANs) are state-of-the-art neural network models used to synthesise images and other data. GANs brought a considerable improvement to the quality of synthetic data, quickly becoming the standard for data-generation tasks. In this work, we summarise the applications of GANs in the field of cardiology, including generation of realistic cardiac images, electrocardiography signals, and synthetic electronic health records. The utility of GAN-generated data is discussed with respect to research, clinical care, and academia. And we present illustrative examples of our GAN-generated cardiac magnetic resonance and echocardiography images, showing the evolution in image quality across 6 different models, which have become almost indistinguishable from real images. Finally, we discuss future applications, such as modality translation or patient trajectory modelling. Moreover, we discuss the pending challenges that GANs need to overcome, namely, their training dynamics, the medical fidelity or the data regulations and ethics questions, to become integrated in cardiology workflows., (Copyright © 2021 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.)

Current brain white matter fiber tracking techniques show a number of problems, including: generating large proportions of streamlines that do not accurately describe the underlying anatomy; extracting streamlines that are not supported by the underlying diffusion signal; and under-representing some fiber populations, among others. In this paper, we describe a novel autoencoder-based learning method to filter streamlines from diffusion MRI tractography, and hence, to obtain more reliable tractograms. Our method, dubbed FINTA (Filtering in Tractography using Autoencoders) uses raw, unlabeled tractograms to train the autoencoder, and to learn a robust representation of brain streamlines. Such an embedding is then used to filter undesired streamline samples using a nearest neighbor algorithm. Our experiments on both synthetic and in vivo human brain diffusion MRI tractography data obtain accuracy scores exceeding the 90% threshold on the test set. Results reveal that FINTA has a superior filtering performance compared to conventional, anatomy-based methods, and the RecoBundles state-of-the-art method. Additionally, we demonstrate that FINTA can be applied to partial tractograms without requiring changes to the framework. We also show that the proposed method generalizes well across different tracking methods and datasets, and shortens significantly the computation time for large (>1 M streamlines) tractograms. Together, this work brings forward a new deep learning framework in tractography based on autoencoders, which offers a flexible and powerful method for white matter filtering and bundling that could enhance tractometry and connectivity analyses., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2021 Elsevier B.V. All rights reserved.)

Diffusion MRI tractography is currently the only non-invasive tool able to assess the white-matter structural connectivity of a brain. Since its inception, it has been widely documented that tractography is prone to producing erroneous tracks while missing true positive connections. Recently, supervised learning algorithms have been proposed to learn the tracking procedure implicitly from data, without relying on anatomical priors. However, these methods rely on curated streamlines that are very hard to obtain. To remove the need for such data but still leverage the expressiveness of neural networks, we introduce Track-To-Learn: A general framework to pose tractography as a deep reinforcement learning problem. Deep reinforcement learning is a type of machine learning that does not depend on ground-truth data but rather on the concept of "reward". We implement and train algorithms to maximize returns from a reward function based on the alignment of streamlines with principal directions extracted from diffusion data. We show competitive results on known data and little loss of performance when generalizing to new, unseen data, compared to prior machine learning-based tractography algorithms. To the best of our knowledge, this is the first successful use of deep reinforcement learning for tractography., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2021 Elsevier B.V. All rights reserved.)

This paper presents a client/server privacy-preserving network in the context of multicentric medical image analysis. Our approach is based on adversarial learning which encodes images to obfuscate the patient identity while preserving enough information for a target task. Our novel architecture is composed of three components: 1) an encoder network which removes identity-specific features from input medical images, 2) a discriminator network that attempts to identify the subject from the encoded images, 3) a medical image analysis network which analyzes the content of the encoded images (segmentation in our case). By simultaneously fooling the discriminator and optimizing the medical analysis network, the encoder learns to remove privacy-specific features while keeping those essentials for the target task. Our approach is illustrated on the problem of segmenting brain MRI from the large-scale Parkinson Progression Marker Initiative (PPMI) dataset. Using longitudinal data from PPMI, we show that the discriminator learns to heavily distort input images while allowing for highly accurate segmentation results. Our results also demonstrate that an encoder trained on the PPMI dataset can be used for segmenting other datasets, without the need for retraining. The code is made available at: https://github.com/bachkimn/Privacy-Net-An-Adversarial-Approach-forIdentity-Obfuscated-Segmentation-of-MedicalImages.

Segmentation of cardiac structures is one of the fundamental steps to estimate volumetric indices of the heart. This step is still performed semiautomatically in clinical routine and is, thus, prone to interobserver and intraobserver variabilities. Recent studies have shown that deep learning has the potential to perform fully automatic segmentation. However, the current best solutions still suffer from a lack of robustness in terms of accuracy and number of outliers. The goal of this work is to introduce a novel network designed to improve the overall segmentation accuracy of left ventricular structures (endocardial and epicardial borders) while enhancing the estimation of the corresponding clinical indices and reducing the number of outliers. This network is based on a multistage framework where both the localization and segmentation steps are optimized jointly through an end-to-end scheme. Results obtained on a large open access data set show that our method outperforms the current best-performing deep learning solution with a lighter architecture and achieved an overall segmentation accuracy lower than the intraobserver variability for the epicardial border (i.e., on average a mean absolute error of 1.5 mm and a Hausdorff distance of 5.1mm) with 11% of outliers. Moreover, we demonstrate that our method can closely reproduce the expert analysis for the end-diastolic and end-systolic left ventricular volumes, with a mean correlation of 0.96 and a mean absolute error of 7.6 ml. Concerning the ejection fraction of the left ventricle, results are more contrasted with a mean correlation coefficient of 0.83 and an absolute mean error of 5.0%, producing scores that are slightly below the intraobserver margin. Based on this observation, areas for improvement are suggested.

Convolutional neural networks (CNN) have had unprecedented success in medical imaging and, in particular, in medical image segmentation. However, despite the fact that segmentation results are closer than ever to the inter-expert variability, CNNs are not immune to producing anatomically inaccurate segmentations, even when built upon a shape prior. In this paper, we present a framework for producing cardiac image segmentation maps that are guaranteed to respect pre-defined anatomical criteria, while remaining within the inter-expert variability. The idea behind our method is to use a well-trained CNN, have it process cardiac images, identify the anatomically implausible results and warp these results toward the closest anatomically valid cardiac shape. This warping procedure is carried out with a constrained variational autoencoder (cVAE) trained to learn a representation of valid cardiac shapes through a smooth, yet constrained, latent space. With this cVAE, we can project any implausible shape into the cardiac latent space and steer it toward the closest correct shape. We tested our framework on short-axis MRI as well as apical two and four-chamber view ultrasound images, two modalities for which cardiac shapes are drastically different. With our method, CNNs can now produce results that are both within the inter-expert variability and always anatomically plausible without having to rely on a shape prior.

Supervised machine learning (ML) algorithms have recently been proposed as an alternative to traditional tractography methods in order to address some of their weaknesses. They can be path-based and local-model-free, and easily incorporate anatomical priors to make contextual and non-local decisions that should help the tracking process. ML-based techniques have thus shown promising reconstructions of larger spatial extent of existing white matter bundles, promising reconstructions of less false positives, and promising robustness to known position and shape biases of current tractography techniques. But as of today, none of these ML-based methods have shown conclusive performances or have been adopted as a de facto solution to tractography. One reason for this might be the lack of well-defined and extensive frameworks to train, evaluate, and compare these methods. In this paper, we describe several datasets and evaluation tools that contain useful features for ML algorithms, along with the various methods proposed in the recent years. We then discuss the strategies that are used to evaluate and compare those methods, as well as their shortcomings. Finally, we describe the particular needs of ML tractography methods and discuss tangible solutions for future works., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Recent evidence shows that neuroinflammation plays a role in many neurological diseases including mild cognitive impairment (MCI) and Alzheimer's disease (AD), and that free water (FW) modeling from clinically acquired diffusion MRI (DTI-like acquisitions) can be sensitive to this phenomenon. This FW index measures the fraction of the diffusion signal explained by isotropically unconstrained water, as estimated from a bi-tensor model. In this study, we developed a simple but powerful whole-brain FW measure designed for easy translation to clinical settings and potential use as a priori outcome measure in clinical trials. These simple FW measures use a "safe" white matter (WM) mask without gray matter (GM)/CSF partial volume contamination ( WM safe ) near ventricles and sulci. We investigated if FW inside the WM safe mask, including and excluding areas of white matter damage such as white matter hyperintensities (WMHs) as shown on T2 FLAIR, computed across the whole white matter could be indicative of diagnostic grouping along the AD continuum. After careful quality control, 81 cognitively normal controls (NC), 103 subjects with MCI and 42 with AD were selected from the ADNIGO and ADNI2 databases. We show that MCI and AD have significantly higher FW measures even after removing all partial volume contamination. We also show, for the first time, that when WMHs are removed from the masks, the significant results are maintained, which demonstrates that the FW measures are not just a byproduct of WMHs. Our new and simple FW measures can be used to increase our understanding of the role of inflammation-associated edema in AD and may aid in the differentiation of healthy subjects from MCI and AD patients., (Copyright © 2019 Dumont, Roy, Jodoin, Morency, Houde, Xie, Bauer, Samad, Van Dijk, Goodman and Descoteaux.)

Delineation of the cardiac structures from 2D echocardiographic images is a common clinical task to establish a diagnosis. Over the past decades, the automation of this task has been the subject of intense research. In this paper, we evaluate how far the state-of-the-art encoder-decoder deep convolutional neural network methods can go at assessing 2D echocardiographic images, i.e., segmenting cardiac structures and estimating clinical indices, on a dataset, especially, designed to answer this objective. We, therefore, introduce the cardiac acquisitions for multi-structure ultrasound segmentation dataset, the largest publicly-available and fully-annotated dataset for the purpose of echocardiographic assessment. The dataset contains two and four-chamber acquisitions from 500 patients with reference measurements from one cardiologist on the full dataset and from three cardiologists on a fold of 50 patients. Results show that encoder-decoder-based architectures outperform state-of-the-art non-deep learning methods and faithfully reproduce the expert analysis for the end-diastolic and end-systolic left ventricular volumes, with a mean correlation of 0.95 and an absolute mean error of 9.5 ml. Concerning the ejection fraction of the left ventricle, results are more contrasted with a mean correlation coefficient of 0.80 and an absolute mean error of 5.6%. Although these results are below the inter-observer scores, they remain slightly worse than the intra-observer's ones. Based on this observation, areas for improvement are defined, which open the door for accurate and fully-automatic analysis of 2D echocardiographic images.

In this paper, we present a novel convolutional neural network architecture to segment images from a series of short-axis cardiac magnetic resonance slices (CMRI). The proposed model is an extension of the U-net that embeds a cardiac shape prior and involves a loss function tailored to the cardiac anatomy. Since the shape prior is computed offline only once, the execution of our model is not limited by its calculation. Our system takes as input raw magnetic resonance images, requires no manual preprocessing or image cropping and is trained to segment the endocardium and epicardium of the left ventricle, the endocardium of the right ventricle, as well as the center of the left ventricle. With its multiresolution grid architecture, the network learns both high and low-level features useful to register the shape prior as well as accurately localize the borders of the cardiac regions. Experimental results obtained on the Automatic Cardiac Diagnostic Challenge - Medical Image Computing and Computer Assisted Intervention (ACDC-MICCAI) 2017 dataset show that our model segments multislices CMRI (left and right ventricle contours) in 0.18 s with an average Dice coefficient of [Formula: see text] and an average 3-D Hausdorff distance of [Formula: see text] mm.

Tissue segmentation and classification in MRI is a challenging task due to a lack of signal intensity standardization. MRI signal is dependent on the acquisition protocol, the coil profile, the scanner type, etc. While we can compute quantitative physical tissue properties independent of the hardware and the sequence parameters, it is still difficult to leverage these physical properties to segment and classify pelvic tissues. The proposed method integrates quantitative MRI values (T1 and T2 relaxation times and pure synthetic weighted images) and machine learning (Support Vector Machine (SVM)) to segment and classify tissues in the pelvic region, i.e.: fat, muscle, prostate, bone marrow, bladder, and air. Twenty-two men with a mean age of 30±14 years were included in this prospective study. The images were acquired with a 3 Tesla MRI scanner. An inversion recovery-prepared turbo spin echo sequence was used to obtain T1-weighted images at different inversion times with a TR of 14000 ms. A 32-echo spin echo sequence was used to obtain the T2-weighted images at different echo times with a TR of 5000 ms. T1 and T2 relaxation times, synthetic T1- and T2-weighted images and anatomical probabilistic maps were calculated and used as input features of a SVM for segmenting and classifying tissues within the pelvic region. The mean SVM classification accuracy across subjects was calculated for the different tissues: prostate (94.2%), fat (96.9%), muscle (95.8%), bone marrow (91%) and bladder (82.1%) indicating an excellent classification performance. However, the segmentation and classification for air (within the rectum) may not always be successful (mean SVM accuracy 47.5%) due to the lack of air data in the training and testing sets. Our findings suggest that SVM can reliably segment and classify tissues in the pelvic region., Competing Interests: The authors have declared that no competing interests exist.

Delineation of the left ventricular cavity, myocardium, and right ventricle from cardiac magnetic resonance images (multi-slice 2-D cine MRI) is a common clinical task to establish diagnosis. The automation of the corresponding tasks has thus been the subject of intense research over the past decades. In this paper, we introduce the "Automatic Cardiac Diagnosis Challenge" dataset (ACDC), the largest publicly available and fully annotated dataset for the purpose of cardiac MRI (CMR) assessment. The dataset contains data from 150 multi-equipments CMRI recordings with reference measurements and classification from two medical experts. The overarching objective of this paper is to measure how far state-of-the-art deep learning methods can go at assessing CMRI, i.e., segmenting the myocardium and the two ventricles as well as classifying pathologies. In the wake of the 2017 MICCAI-ACDC challenge, we report results from deep learning methods provided by nine research groups for the segmentation task and four groups for the classification task. Results show that the best methods faithfully reproduce the expert analysis, leading to a mean value of 0.97 correlation score for the automatic extraction of clinical indices and an accuracy of 0.96 for automatic diagnosis. These results clearly open the door to highly accurate and fully automatic analysis of cardiac CMRI. We also identify scenarios for which deep learning methods are still failing. Both the dataset and detailed results are publicly available online, while the platform will remain open for new submissions.

The ability to train on a large dataset of labeled samples is critical to the success of deep learning in many domains. In this paper, we focus on motor vehicle classification and localization from a single video frame and introduce the "MIOvision Traffic Camera Dataset" (MIO-TCD) in this context. MIO-TCD is the largest dataset for motorized traffic analysis to date. It includes 11 traffic object classes such as cars, trucks, buses, motorcycles, bicycles, pedestrians. It contains 786,702 annotated images acquired at different times of the day and different periods of the year by hundreds of traffic surveillance cameras deployed across Canada and the United States. The dataset consists of two parts: a "localization dataset", containing 137,743 full video frames with bounding boxes around traffic objects, and a "classification dataset", containing 648,959 crops of traffic objects from the 11 classes. We also report results from the 2017 CVPR MIO-TCD Challenge, that leveraged this dataset, and compare them with results for state-of-the-art deep learning architectures. These results demonstrate the viability of deep learning methods for vehicle localization and classification from a single video frame in real-life traffic scenarios. The topperforming methods achieve both accuracy and Kappa score above 96% on the classification dataset and mean-average precision of 77% on the localization dataset. We also identify scenarios in which state-of-the-art methods still fail and we suggest avenues to address these challenges. Both the dataset and detailed results are publicly available on-line [1].

Background: In temporal lobe epilepsy (TLE), advanced neuroimaging techniques reveal anomalies extending beyond the temporal lobe such as thinning of fronto-central cortices. Operculo-insular epilepsy (OIE) is an under-recognized and poorly characterized condition with the potential of mimicking TLE. In this work, we investigated insular and extra-insular cortical thickness (CT) changes in OIE., Methods: All participants (14 patients with refractory OIE, 9 age- and sex-matched patients with refractory TLE and 26 healthy controls) underwent a T1-weighted acquisition on a 3 T MRI. Anatomical images were processed with Advanced Normalization Tools. Between-group analysis of CT was performed using a two-sided t -test (threshold of p < 0.05 after correction for multiple comparisons; cut-off threshold of 250 voxels) between (i) patients with OIE vs TLE, and (ii) patients with OIE vs healthy controls., Results: Significant widespread thinning was observed in OIE patients as compared with healthy controls mainly in the ipsilateral insula, peri-rolandic region, orbito-frontal area, mesiotemporal structures and lateral temporal neocortex. Contralateral cortical shrinkage followed a similar albeit milder and less diffuse pattern.The CT of OIE patients was equal or reduced relative to the TLE group for every cortical region analyzed. Thinning was observed diffusely in OIE patients, predominantly inboth insulae and the ipsilateral occipito-temporal area., Conclusion: Our results reveal structural anomalies extending beyond the operculo-insular area in OIE.

Scene background initialization is the process by which a method tries to recover the background image of a video without foreground objects in it. Having a clear understanding about which approach is more robust and/or more suited to a given scenario is of great interest to many end users or practitioners. The aim of this paper is to provide an extensive survey of scene background initialization methods as well as a novel benchmarking framework. The proposed framework involves several evaluation metrics and state-of-the-art methods, as well as the largest video data set ever made for this purpose. The data set consists of several camera-captured videos that: 1) span categories focused on various background initialization challenges; 2) are obtained with different cameras of different lengths, frame rates, spatial resolutions, lighting conditions, and levels of compression; and 3) contain indoor and outdoor scenes. The wide variety of our data set prevents our analysis from favoring a certain family of background initialization methods over others. Our evaluation framework allows us to quantitatively identify solved and unsolved issues related to scene background initialization. We also identify scenarios for which state-of-the-art methods systematically fail.

In this work, we propose a diffusion MRI protocol for mining Parkinson's disease diffusion MRI datasets and recover robust disease-specific biomarkers. Using advanced high angular resolution diffusion imaging (HARDI) crossing fiber modeling and tractography robust to partial volume effects, we automatically dissected 50 white matter (WM) fascicles. These fascicles connect deep nuclei (thalamus, putamen, pallidum) to different cortical functional areas (associative, motor, sensorimotor, limbic), basal forebrain and substantia nigra. Then, among these 50 candidate WM fascicles, only the ones that passed a test-retest reproducibility procedure qualified for further tractometry analysis. Leveraging the unique 2-timepoints test-retest Parkinson's Progression Markers Initiative (PPMI) dataset of over 600 subjects, we found statistically significant differences in tract profiles along the subcortico-cortical pathways between Parkinson's disease patients and healthy controls. In particular, significant increases in FA, apparent fiber density, tract-density and generalized FA were detected in some locations of the nigro-subthalamo-putaminal-thalamo-cortical pathway. This connection is one of the major motor circuits balancing the coordination of motor output. Detailed and quantifiable knowledge on WM fascicles in these areas is thus essential to improve the quality and outcome of Deep Brain Stimulation, and to target new WM locations for investigation.

In this paper, we present a fast and interactive graph cut method for 3D segmentation of the endocardial wall of the left ventricle (LV) adapted to work on two of the most widely used modalities: magnetic resonance imaging (MRI) and echocardiography. Our method accounts for the fundamentally different nature of both modalities: 3D echocardiographic images have a low contrast, a poor signal-to-noise ratio and frequent signal drop, while MR images are more detailed but also cluttered and contain highly anisotropic voxels. The main characteristic of our method is to work in a 3D Bezier coordinate system instead of the original Euclidean space. This comes with several advantages, including an implicit shape prior and a result guarantied not to have any holes in it. The proposed method is made of 4 steps. First, a 3D sampling of the LV cavity is made based on a Bezier coordinate system. This allows to warp the input 3D image to a Bezier space in which a plane corresponds to an anatomically plausible 3D Euclidean bullet shape. Second, a 3D graph is built and an energy term (which is based on the image gradient and a 3D probability map) is assigned to each edge of the graph, some of which being given an infinite energy to ensure the resulting 3D structure passes through key anatomical points. Third, a max-flow min-cut procedure is executed on the energy graph to delineate the endocardial surface. And fourth, the resulting surface is projected back to the Euclidean space where a post-processing convex hull algorithm is applied on every short axis slice to remove local concavities. Results obtained on two datasets reveal that our method takes between 2 and 5s to segment a 3D volume, it has better results overall than most state-of-the-art methods on the CETUS echocardiographic dataset and is statistically as good as a human operator on MR images., (Copyright © 2017 Elsevier Ltd. All rights reserved.)

In this paper, we present a fully automatic brain tumor segmentation method based on Deep Neural Networks (DNNs). The proposed networks are tailored to glioblastomas (both low and high grade) pictured in MR images. By their very nature, these tumors can appear anywhere in the brain and have almost any kind of shape, size, and contrast. These reasons motivate our exploration of a machine learning solution that exploits a flexible, high capacity DNN while being extremely efficient. Here, we give a description of different model choices that we've found to be necessary for obtaining competitive performance. We explore in particular different architectures based on Convolutional Neural Networks (CNN), i.e. DNNs specifically adapted to image data. We present a novel CNN architecture which differs from those traditionally used in computer vision. Our CNN exploits both local features as well as more global contextual features simultaneously. Also, different from most traditional uses of CNNs, our networks use a final layer that is a convolutional implementation of a fully connected layer which allows a 40 fold speed up. We also describe a 2-phase training procedure that allows us to tackle difficulties related to the imbalance of tumor labels. Finally, we explore a cascade architecture in which the output of a basic CNN is treated as an additional source of information for a subsequent CNN. Results reported on the 2013 BRATS test data-set reveal that our architecture improves over the currently published state-of-the-art while being over 30 times faster., (Copyright © 2016 Elsevier B.V. All rights reserved.)

Ischemic stroke is the most common cerebrovascular disease, and its diagnosis, treatment, and study relies on non-invasive imaging. Algorithms for stroke lesion segmentation from magnetic resonance imaging (MRI) volumes are intensely researched, but the reported results are largely incomparable due to different datasets and evaluation schemes. We approached this urgent problem of comparability with the Ischemic Stroke Lesion Segmentation (ISLES) challenge organized in conjunction with the MICCAI 2015 conference. In this paper we propose a common evaluation framework, describe the publicly available datasets, and present the results of the two sub-challenges: Sub-Acute Stroke Lesion Segmentation (SISS) and Stroke Perfusion Estimation (SPES). A total of 16 research groups participated with a wide range of state-of-the-art automatic segmentation algorithms. A thorough analysis of the obtained data enables a critical evaluation of the current state-of-the-art, recommendations for further developments, and the identification of remaining challenges. The segmentation of acute perfusion lesions addressed in SPES was found to be feasible. However, algorithms applied to sub-acute lesion segmentation in SISS still lack accuracy. Overall, no algorithmic characteristic of any method was found to perform superior to the others. Instead, the characteristics of stroke lesion appearances, their evolution, and the observed challenges should be studied in detail. The annotated ISLES image datasets continue to be publicly available through an online evaluation system to serve as an ongoing benchmarking resource (www.isles-challenge.org)., (Copyright © 2016 Elsevier B.V. All rights reserved.)

Purpose: In this paper, we investigate a framework for interactive brain tumor segmentation which, at its core, treats the problem of interactive brain tumor segmentation as a machine learning problem., Methods: This method has an advantage over typical machine learning methods for this task where generalization is made across brains. The problem with these methods is that they need to deal with intensity bias correction and other MRI-specific noise. In this paper, we avoid these issues by approaching the problem as one of within brain generalization. Specifically, we propose a semi-automatic method that segments a brain tumor by training and generalizing within that brain only, based on some minimum user interaction., Conclusion: We investigate how adding spatial feature coordinates (i.e., i, j, k) to the intensity features can significantly improve the performance of different classification methods such as SVM, kNN and random forests. This would only be possible within an interactive framework. We also investigate the use of a more appropriate kernel and the adaptation of hyper-parameters specifically for each brain., Results: As a result of these experiments, we obtain an interactive method whose results reported on the MICCAI-BRATS 2013 dataset are the second most accurate compared to published methods, while using significantly less memory and processing power than most state-of-the-art methods.

Real-time 3D Echocardiography (RT3DE) has been proven to be an accurate tool for left ventricular (LV) volume assessment. However, identification of the LV endocardium remains a challenging task, mainly because of the low tissue/blood contrast of the images combined with typical artifacts. Several semi and fully automatic algorithms have been proposed for segmenting the endocardium in RT3DE data in order to extract relevant clinical indices, but a systematic and fair comparison between such methods has so far been impossible due to the lack of a publicly available common database. Here, we introduce a standardized evaluation framework to reliably evaluate and compare the performance of the algorithms developed to segment the LV border in RT3DE. A database consisting of 45 multivendor cardiac ultrasound recordings acquired at different centers with corresponding reference measurements from three experts are made available. The algorithms from nine research groups were quantitatively evaluated and compared using the proposed online platform. The results showed that the best methods produce promising results with respect to the experts' measurements for the extraction of clinical indices, and that they offer good segmentation precision in terms of mean distance error in the context of the experts' variability range. The platform remains open for new submissions.

Single-photon emission computed tomography (SPECT) imaging of the heart is helpful to quantify the left-ventricular ejection fraction and study myocardial perfusion scans. However, these evaluations require a 3-D segmentation of the left-ventricular wall on each phase of the cardiac cycle. This paper presents a fast and interactive graph cut method for 3-D segmentation of the left ventricle (LV) of rats in SPECT images. The method is carried out in three steps. First, 3-D sampling of the LV cavity is made in a spherical-cylindrical coordinate system. Then, a graph-cut-based energy minimization procedure provides delineation of the myocardium centerline surface. From there, it is possible to outline the epicardial and endocardial boundaries by considering the second statistical moment of the SPECT images. An important aspect of our method is to always produce anatomically coherent U-shape results. It also relies on only two intuitive parameters regulating the smoothness and the thickness of the segmentation result. Results show not only that our method is statistically as accurate as human experts, but it is one order of magnitude faster than a state-of-the-art method with a processing time of at most 2 s on a 4-D cardiac image after having determined the LV orientation.

A single diffusion MRI streamline fiber tracking dataset may contain hundreds of thousands, and often millions of streamlines and can take up to several gigabytes of memory. This amount of data is not only heavy to compute, but also difficult to visualize and hard to store on disk (especially when dealing with a collection of brains). These problems call for a fiber-specific compression format that simplifies its manipulation. As of today, no fiber compression format has yet been adopted and the need for it is now becoming an issue for future connectomics research. In this work, we propose a new compression format, .zfib, for streamline tractography datasets reconstructed from diffusion magnetic resonance imaging (dMRI). Tracts contain a large amount of redundant information and are relatively smooth. Hence, they are highly compressible. The proposed method is a processing pipeline containing a linearization, a quantization and an encoding step. Our pipeline is tested and validated under a wide range of DTI and HARDI tractography configurations (step size, streamline number, deterministic and probabilistic tracking) and compression options. Similar to JPEG, the user has one parameter to select: a worst-case maximum tolerance error in millimeter (mm). Overall, we find a compression factor of more than 96% for a maximum error of 0.1mm without any perceptual change or change of diffusion statistics (mean fractional anisotropy and mean diffusivity) along bundles. This opens new opportunities for connectomics and tractometry applications., (Copyright © 2014 Elsevier Inc. All rights reserved.)

Change detection is one of the most commonly encountered low-level tasks in computer vision and video processing. A plethora of algorithms have been developed to date, yet no widely accepted, realistic, large-scale video data set exists for benchmarking different methods. Presented here is a unique change detection video data set consisting of nearly 90 000 frames in 31 video sequences representing six categories selected to cover a wide range of challenges in two modalities (color and thermal infrared). A distinguishing characteristic of this benchmark video data set is that each frame is meticulously annotated by hand for ground-truth foreground, background, and shadow area boundaries-an effort that goes much beyond a simple binary label denoting the presence of change. This enables objective and precise quantitative comparison and ranking of video-based change detection algorithms. This paper discusses various aspects of the new data set, quantitative performance metrics used, and comparative results for over two dozen change detection algorithms. It draws important conclusions on solved and remaining issues in change detection, and describes future challenges for the scientific community. The data set, evaluation tools, and algorithm rankings are available to the public on a website and will be updated with feedback from academia and industry in the future.

Background subtraction has been a driving engine for many computer vision and video analytics tasks. Although its many variants exist, they all share the underlying assumption that photometric scene properties are either static or exhibit temporal stationarity. While this works in many applications, the model fails when one is interested in discovering changes in scene dynamics instead of changes in scene's photometric properties; the detection of unusual pedestrian or motor traffic patterns are but two examples. We propose a new model and computational framework that assume the dynamics of a scene, not its photometry, to be stationary, i.e., a dynamic background serves as the reference for the dynamics of an observed scene. Central to our approach is the concept of an event, which we define as short-term scene dynamics captured over a time window at a specific spatial location in the camera field of view. Unlike in our earlier work, we compute events by time-aggregating vector object descriptors that can combine multiple features, such as object size, direction of movement, speed, etc. We characterize events probabilistically, but use low-memory, low-complexity surrogates in a practical implementation. Using these surrogates amounts to behavior subtraction, a new algorithm for effective and efficient temporal anomaly detection and localization. Behavior subtraction is resilient to spurious background motion, such as due to camera jitter, and is content-blind, i.e., it works equally well on humans, cars, animals, and other objects in both uncluttered and highly cluttered scenes. Clearly, treating video as a collection of events rather than colored pixels opens new possibilities for video analytics.

We designed a generic method for segmenting the aneurismal sac of an abdominal aortic aneurysm (AAA) both from multi-slice MR and CT-scan examinations. It is a semi-automatic method requiring little human intervention and based on graph cut theory to segment the lumen interface and the aortic wall of AAAs. Our segmentation method works independently on MRI and CT-scan volumes and has been tested on a 44 patient dataset and 10 synthetic images. Segmentation and maximum diameter estimation were compared to manual tracing from 4 experts. An inter-observer study was performed in order to measure the variability range of a human observer. Based on three metrics (the maximum aortic diameter, the volume overlap and the Hausdorff distance) the variability of the results obtained by our method is shown to be similar to that of a human operator, both for the lumen interface and the aortic wall. As will be shown, the average distance obtained with our method is less than one standard deviation away from each expert, both for healthy subjects and for patients with AAA. Our semi-automatic method provides reliable contours of the abdominal aorta from CT-scan or MRI, allowing rapid and reproducible evaluations of AAA., (Copyright © 2011 Elsevier Ltd. All rights reserved.)

We present a method for estimating the amount of noise and blur in a distorted image. Our method is based on the multiscale structural similarity (MS-SSIM) framework that, although designed to measure image quality, is used to estimate the amount of blur and noise in a degraded image given a reference image. We show that there exists a bijective mapping between the 2-D noise/blur space and the 3-D MS-SSIM space, which allows to recover distortion parameters. That mapping allows to formulate the multidistortion-estimation problem as a classical optimization problem. Various search strategies such as Newton, simplex, NewUOA, and brute-force search are presented and compared. We also show that a bicubic patch can be used to approximate the bijective mapping between the noise/blur space and the 3-D MS-SSIM space. Interestingly, the use of such a patch reduces the processing time by a factor of 40 without significantly reducing precision. Based on quantitative results, we show that the amount of different types of blur and noise in a distorted image can be recovered with accuracy of roughly 2% and 8%, respectively. Our methods are compared with four state-of-the-art noise- and blur-estimation techniques., (© 2011 IEEE)

In this paper, we consider the problem of finding correspondences between distributed cameras that have partially overlapping field of views. When multiple cameras with adaptable orientations and zooms are deployed, as in many wide area surveillance applications, identifying correspondence between different activities becomes a fundamental issue. We propose a correspondence method based upon activity features that, unlike photometric features, have certain geometry independence properties. The proposed method is robust to pose, illumination and geometric effects, unsupervised (does not require any calibration objects). In addition, these features are amenable to low communication bandwidth and distributed network applications. We present quantitative and qualitative results with synthetic and real life examples, and compare the proposed method with scale invariant feature transform (SIFT) based method. We show that our method significantly outperforms the SIFT method when cameras have significantly different orientations. We then describe extensions of our method in a number of directions including topology reconstruction, camera calibration, and distributed anomaly detection.

In this paper, we put forward a novel fusion framework that mixes together label fields instead of observation data as is usually the case. Our framework takes as input two label fields: a quickly estimated and to-be-refined segmentation map and a spatial region map that exhibits the shape of the main objects of the scene. These two label fields are fused together with a global energy function that is minimized with a deterministic iterative conditional mode algorithm. As explained in the paper, the energy function may implement a pure fusion strategy or a fusion-reaction function. In the latter case, a data-related term is used to make the optimization problem well posed. We believe that the conceptual simplicity, the small number of parameters, the use of a simple and fast deterministic optimizer that admits a natural implementation on a parallel architecture are among the main advantages of our approach. Our fusion framework is adapted to various computer vision applications among which are motion segmentation, motion estimation and occlusion detection.