Nathan Painchaud

Nathan Painchaud

Postdoctoral researcher

Posts by Collection

portfolio

publications

Cardiac MRI Segmentation with Strong Anatomical Guarantees

Published in Medical Image Computing and Computer Assisted Intervention (MICCAI), 2019

Recent publications have shown that the segmentation accuracy of modern-day convolutional neural networks (CNN) applied on cardiac MRI can reach the inter-expert variability, a great achievement in this area of research. However, despite these successes, CNNs still produce anatomically inaccurate segmentations as they provide no guarantee on the anatomical plausibility of their outcome, even when using a shape prior. In this paper, we propose a cardiac MRI segmentation method which always produces anatomically plausible results. At the core of the method is an adversarial variational autoencoder (aVAE) whose latent space encodes a smooth manifold on which lies a large spectrum of valid cardiac shapes. This aVAE is used to automatically warp anatomically inaccurate cardiac shapes towards a close but correct shape. Our method can accommodate any cardiac segmentation method and convert its anatomically implausible results to plausible ones without affecting its overall geometric and clinical metrics. With our method, CNNs can now produce results that are both within the inter-expert variability and always anatomically plausible.

Recommended citation: N. Painchaud, Y. Skandarani, T. Judge, O. Bernard, A. Lalande, and P.-M. Jodoin, "Cardiac MRI Segmentation with Strong Anatomical Guarantees," in proc. Medical Image Computing and Computer Assisted Intervention (MICCAI), 2019, pp. 632–640. https://doi.org/10.1007/978-3-030-32245-8_70

On the effectiveness of GAN generated cardiac MRIs for segmentation

Published in Medical Imaging with Deep Learning (MIDL), 2020

In this work, we propose a Variational Autoencoder (VAE) - Generative Adversarial Networks (GAN) model that can produce highly realistic MRI together with its pixel accurate groundtruth for the application of cine-MR image cardiac segmentation. On one side of our model is a Variational Autoencoder (VAE) trained to learn the latent representations of cardiac shapes. On the other side is a GAN that uses ”SPatially-Adaptive (DE)Normalization” (SPADE) modules to generate realistic MR images tailored to a given anatomical map. At test time, the sampling of the VAE latent space allows to generate an arbitrary large number of cardiac shapes, which are fed to the GAN that subsequently generates MR images whose cardiac structure fits that of the cardiac shapes. In other words, our system can generate a large volume of realistic yet labeled cardiac MR images. We show that segmentation with CNNs trained with our synthetic annotated images gets competitive results compared to traditional techniques. We also show that combining data augmentation with our GAN-generated images lead to an improvement in the Dice score of up to 12 percent while allowing for better generalization capabilities on other datasets.

Recommended citation: Y. Skandarani, N. Painchaud, P.-M. Jodoin, and A. Lalande, "On the effectiveness of GAN generated cardiac MRIs for segmentation," presented at Medical Imaging with Deep Learning (MIDL), 2020. https://openreview.net/forum?id=f9Pl3Qj3_Q

Cardiac Segmentation with Strong Anatomical Guarantees

Published in IEEE Transactions on Medical Imaging (IEEE TMI), 2020

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.

Recommended citation: N. Painchaud, Y. Skandarani, T. Judge, O. Bernard, A. Lalande, and P.-M. Jodoin, "Cardiac Segmentation with Strong Anatomical Guarantees," IEEE Transactions on Medical Imaging, vol. 39, no. 11, pp. 3703–3713, Nov. 2020. https://doi.org/10.1109/TMI.2020.3003240

Echocardiography Segmentation With Enforced Temporal Consistency

Published in IEEE Transactions on Medical Imaging (IEEE TMI), 2022

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.

Recommended citation: N. Painchaud, N. Duchateau, O. Bernard, and P.-M. Jodoin, "Echocardiography Segmentation With Enforced Temporal Consistency," IEEE Transactions on Medical Imaging, vol. 41, no. 10, pp. 2867–2878, Oct. 2022. https://doi.org/10.1109/TMI.2022.3173669

Neural Teleportation

Published in Mathematics, 2023

In this paper, we explore a process called neural teleportation, a mathematical consequence of applying quiver representation theory to neural networks. Neural teleportation teleports a network to a new position in the weight space and preserves its function. This phenomenon comes directly from the definitions of representation theory applied to neural networks and it turns out to be a very simple operation that has remarkable properties. We shed light on the surprising and counter-intuitive consequences neural teleportation has on the loss landscape. In particular, we show that teleportation can be used to explore loss level curves, that it changes the local loss landscape, sharpens global minima and boosts back-propagated gradients at any moment during the learning process.

Recommended citation: M. Armenta, T. Judge, N. Painchaud, Y. Skandarani, C. Lemaire, G. G. Sanchez, P. Spino, and P.-M. Jodoin, "Neural Teleportation," Mathematics, vol. 11, no. 2, pp. 480, Jan. 2023. https://doi.org/10.3390/math11020480

Extraction of Volumetric Indices from Echocardiography: Which Deep Learning Solution for Clinical Use?

Published in Functional Imaging and Modeling of the Heart (FIMH), 2023

Deep learning-based methods have spearheaded the automatic analysis of echocardiographic images, taking advantage of the publication of multiple open access datasets annotated by experts (CAMUS being one of the largest public databases). However, these models are still considered unreliable by clinicians due to unresolved issues concerning i) the temporal consistency of their predictions, and ii) their ability to generalize across datasets. In this context, we propose a comprehensive comparison between the current best performing methods in medical/echocardiographic image segmentation, with a particular focus on temporal consistency and cross-dataset aspects. We introduce a new private dataset, named CARDINAL, of apical two-chamber and apical four-chamber sequences, with reference segmentation over the full cardiac cycle. We show that the proposed 3D nnU-Net outperforms alternative 2D and recurrent segmentation methods. We also report that the best models trained on CARDINAL, when tested on CAMUS without any fine-tuning, still manage to perform competitively with respect to prior methods. Overall, the experimental results suggest that with sufficient training data, 3D nnU-Net could become the first automated tool to finally meet the standards of an everyday clinical device.

Recommended citation: H. J. Ling, N. Painchaud, P.-Y. Courand, P.-M. Jodoin, D. Garcia, and O. Bernard, "Extraction of Volumetric Indices from Echocardiography: Which Deep Learning Solution for Clinical Use?," in proc. Functional Imaging and Modeling of the Heart (FIMH), 2023, pp. 245-254. https://doi.org/10.1007/978-3-031-35302-4_25

talks

teaching

Functional programming

Undergraduate course, Université de Sherbrooke, Aug 2017 - Dec 2017

Formalize the notions of procedural abstraction and data abstraction in the context of functional programming.

Neural networks

Graduate course, Université de Sherbrooke, Jan 2020 - Apr 2020

Be familiar with and understand several types of neural networks. Know how to implement them, train them and analyze their performance. Know how to read, understand, synthesize and present scientific papers on neural networks. Be able to reproduce the results of a scientific article or design a new neural network and evaluate its performance.

Concurrent processes and parallelism

Undergraduate course, Université de Sherbrooke, Jan 2021 - Apr 2021

Become familiar with the concepts of concurrent programming. Learn to solve problems using concurrent programming.

Analysis and programming

Undergraduate course, Université de Sherbrooke, Aug 2018 - Dec 2021

Knowing how to analyze a problem, having high standards for the quality of programs, being able to systematically develop good quality programs in the framework of sequential procedural programming.

Compilers and interpreters

Undergraduate course, Université Laval, Aug 2021 - Dec 2021

General compilation process. Lexical analysis and finite automata. Predictive top-down parsing and context-free grammars. Intermediate representation. Attribute grammars. Typing. Runtime infrastructure. Memory allocation. Code generation. Optimization.

Data structures

Undergraduate course, Université de Sherbrooke, Jan 2018 - Dec 2021

Formalize data structures (stacks, lists, trees, etc.); compare and choose the most suitable implementations of the structures according to the problem at hand; put into practice the notions of module and abstract type.