arXiv subjects

Puneet Sharma

Publications and source records attributed to Puneet Sharma.

At least 37 records · Page 2Linked to original sources

arXiv2024

EchoApex: A General-Purpose Vision Foundation Model for Echocardiography

Quantitative evaluation of echocardiography is essential for precise assessment of cardiac condition, monitoring disease progression, and guiding treatment decisions. The diverse nature of echo images, including variations in probe types, manufacturers, and pathologies, poses challenges for developing artificial intelligent models that can generalize across different clinical practice. We introduce EchoApex, the first general-purpose vision foundation model echocardiography with applications on a variety of clinical practice. Leveraging self-supervised learning, EchoApex is pretrained on over 20 million echo images from 11 clinical centres. By incorporating task-specific decoders and adapter modules, we demonstrate the effectiveness of EchoApex on 4 different kind of clinical applications with 28 sub-tasks, including view classification, interactive structure segmentation, left ventricle hypertrophy detection and automated ejection fraction estimation from view sequences. Compared to state-of-the-art task-specific models, EchoApex attains improved performance with a unified image encoding architecture, demonstrating the benefits of model pretraining at scale with in-domain data. Furthermore, EchoApex illustrates the potential for developing a general-purpose vision foundation model tailored specifically for echocardiography, capable of addressing a diverse range of clinical applications with high efficiency and efficacy.

cs.CV↗

arXiv2024

Towards a vision foundation model for comprehensive assessment of Cardiac MRI

Cardiac magnetic resonance imaging (CMR), considered the gold standard for noninvasive cardiac assessment, is a diverse and complex modality requiring a wide variety of image processing tasks for comprehensive assessment of cardiac morphology and function. Advances in deep learning have enabled the development of state-of-the-art (SoTA) models for these tasks. However, model training is challenging due to data and label scarcity, especially in the less common imaging sequences. Moreover, each model is often trained for a specific task, with no connection between related tasks. In this work, we introduce a vision foundation model trained for CMR assessment, that is trained in a self-supervised fashion on 36 million CMR images. We then finetune the model in supervised way for 9 clinical tasks typical to a CMR workflow, across classification, segmentation, landmark localization, and pathology detection. We demonstrate improved accuracy and robustness across all tasks, over a range of available labeled dataset sizes. We also demonstrate improved few-shot learning with fewer labeled samples, a common challenge in medical image analyses. We achieve an out-of-box performance comparable to SoTA for most clinical tasks. The proposed method thus presents a resource-efficient, unified framework for CMR assessment, with the potential to accelerate the development of deep learning-based solutions for image analysis tasks, even with few annotated data available.

eess.IV↗

arXiv2024

Goal-conditioned reinforcement learning for ultrasound navigation guidance

Transesophageal echocardiography (TEE) plays a pivotal role in cardiology for diagnostic and interventional procedures. However, using it effectively requires extensive training due to the intricate nature of image acquisition and interpretation. To enhance the efficiency of novice sonographers and reduce variability in scan acquisitions, we propose a novel ultrasound (US) navigation assistance method based on contrastive learning as goal-conditioned reinforcement learning (GCRL). We augment the previous framework using a novel contrastive patient batching method (CPB) and a data-augmented contrastive loss, both of which we demonstrate are essential to ensure generalization to anatomical variations across patients. The proposed framework enables navigation to both standard diagnostic as well as intricate interventional views with a single model. Our method was developed with a large dataset of 789 patients and obtained an average error of 6.56 mm in position and 9.36 degrees in angle on a testing dataset of 140 patients, which is competitive or superior to models trained on individual views. Furthermore, we quantitatively validate our method's ability to navigate to interventional views such as the Left Atrial Appendage (LAA) view used in LAA closure. Our approach holds promise in providing valuable guidance during transesophageal ultrasound examinations, contributing to the advancement of skill acquisition for cardiac ultrasound practitioners.

cs.CV↗

arXiv2024

DCSM 2.0: Deep Conditional Shape Models for Data Efficient Segmentation

Segmentation is often the first step in many medical image analyses workflows. Deep learning approaches, while giving state-of-the-art accuracies, are data intensive and do not scale well to low data regimes. We introduce Deep Conditional Shape Models 2.0, which uses an edge detector, along with an implicit shape function conditioned on edge maps, to leverage cross-modality shape information. The shape function is trained exclusively on a source domain (contrasted CT) and applied to the target domain of interest (3D echocardiography). We demonstrate data efficiency in the target domain by varying the amounts of training data used in the edge detection stage. We observe that DCSM 2.0 outperforms the baseline at all data levels in terms of Hausdorff distances, and while using 50% or less of the training data in terms of average mesh distance, and at 10% or less of the data with the dice coefficient. The method scales well to low data regimes, with gains of up to 5% in dice coefficient, 2.58 mm in average surface distance and 21.02 mm in Hausdorff distance when using just 2% (22 volumes) of the training data.

eess.IV↗

arXiv2024

Anatomically aware dual-hop learning for pulmonary embolism detection in CT pulmonary angiograms

Pulmonary Embolisms (PE) represent a leading cause of cardiovascular death. While medical imaging, through computed tomographic pulmonary angiography (CTPA), represents the gold standard for PE diagnosis, it is still susceptible to misdiagnosis or significant diagnosis delays, which may be fatal for critical cases. Despite the recently demonstrated power of deep learning to bring a significant boost in performance in a wide range of medical imaging tasks, there are still very few published researches on automatic pulmonary embolism detection. Herein we introduce a deep learning based approach, which efficiently combines computer vision and deep neural networks for pulmonary embolism detection in CTPA. Our method features novel improvements along three orthogonal axes: 1) automatic detection of anatomical structures; 2) anatomical aware pretraining, and 3) a dual-hop deep neural net for PE detection. We obtain state-of-the-art results on the publicly available multicenter large-scale RSNA dataset.

eess.IV↗

arXiv2024

Self-Supervised Learning for Interventional Image Analytics: Towards Robust Device Trackers

An accurate detection and tracking of devices such as guiding catheters in live X-ray image acquisitions is an essential prerequisite for endovascular cardiac interventions. This information is leveraged for procedural guidance, e.g., directing stent placements. To ensure procedural safety and efficacy, there is a need for high robustness no failures during tracking. To achieve that, one needs to efficiently tackle challenges, such as: device obscuration by contrast agent or other external devices or wires, changes in field-of-view or acquisition angle, as well as the continuous movement due to cardiac and respiratory motion. To overcome the aforementioned challenges, we propose a novel approach to learn spatio-temporal features from a very large data cohort of over 16 million interventional X-ray frames using self-supervision for image sequence data. Our approach is based on a masked image modeling technique that leverages frame interpolation based reconstruction to learn fine inter-frame temporal correspondences. The features encoded in the resulting model are fine-tuned downstream. Our approach achieves state-of-the-art performance and in particular robustness compared to ultra optimized reference solutions (that use multi-stage feature fusion, multi-task and flow regularization). The experiments show that our method achieves 66.31% reduction in maximum tracking error against reference solutions (23.20% when flow regularization is used); achieving a success score of 97.95% at a 3x faster inference speed of 42 frames-per-second (on GPU). The results encourage the use of our approach in various other tasks within interventional image analytics that require effective understanding of spatio-temporal semantics.

cs.CV↗

arXiv2023

Towards Establishing Dense Correspondence on Multiview Coronary Angiography: From Point-to-Point to Curve-to-Curve Query Matching

Coronary angiography is the gold standard imaging technique for studying and diagnosing coronary artery disease. However, the resulting 2D X-ray projections lose 3D information and exhibit visual ambiguities. In this work, we aim to establish dense correspondence in multi-view angiography, serving as a fundamental basis for various clinical applications and downstream tasks. To overcome the challenge of unavailable annotated data, we designed a data simulation pipeline using 3D Coronary Computed Tomography Angiography (CCTA). We formulated the problem of dense correspondence estimation as a query matching task over all points of interest in the given views. We established point-to-point query matching and advanced it to curve-to-curve correspondence, significantly reducing errors by minimizing ambiguity and improving topological awareness. The method was evaluated on a set of 1260 image pairs from different views across 8 clinically relevant angulation groups, demonstrating compelling results and indicating the feasibility of establishing dense correspondence in multi-view angiography.

cs.CV↗

arXiv2023

Transitivity And Related Notions For Graph Induced Symbolic Systems

In this paper, we investigate the dynamical behavior of a two dimensional shift $X_G$ (generated by a two dimensional graph $G=(\mathcal{H},\mathcal{V})$) using the adjacency matrices of the generating graph $G$. In particular, we investigate properties such as transitivity, directional transitivity, weak mixing, directional weak mixing and mixing for the shift space $X_G$. We prove that if $(HV)_{ij}\neq 0 \Leftrightarrow (VH)_{ij}\neq 0$ (for all $i,j$), while doubly transitivity (weak mixing) of $X_H$ (or $X_V$) ensures the same for two dimensional shift generated by the graph $G$, directional transitivity (in the direction $(r,s)$) can be characterized through the block representation of $H^rV^s$. We provide necessary and sufficient criteria to establish horizontal (vertical) transitivity for the shift space $X_G$. We also provide examples to establish the necessity of the conditions imposed. Finally, we investigate the decomposability of a given graph into product of graphs with reduced complexity.

math.DS↗

arXiv2023

Bi-modal public transit system for Berlin and Brandenburg

Bi-modla public transport system consisting of a rail-bound line service and a fleet of on-demand shuttles providing connections to the line service stops is studied for the states of Berlin and Brandenburg for user adoption of x = 1% and x = 10% of the total population.

physics.soc-ph↗

arXiv2023

AI-based, automated chamber volumetry from gated, non-contrast CT

Background: Accurate chamber volumetry from gated, non-contrast cardiac CT (NCCT) scans can be useful for potential screening of heart failure. Objectives: To validate a new, fully automated, AI-based method for cardiac volume and myocardial mass quantification from NCCT scans compared to contrasted CT Angiography (CCTA). Methods: Of a retrospectively collected cohort of 1051 consecutive patients, 420 patients had both NCCT and CCTA scans at mid-diastolic phase, excluding patients with cardiac devices. Ground truth values were obtained from the CCTA scans. Results: The NCCT volume computation shows good agreement with ground truth values. Volume differences [95% CI ] and correlation coefficients were: -9.6 [-45; 26] mL, r = 0.98 for LV Total, -5.4 [-24; 13] mL, r = 0.95 for LA, -8.7 [-45; 28] mL, r = 0.94 for RV, -5.2 [-27; 17] mL, r = 0.92 for RA, -3.2 [-42; 36] mL, r = 0.91 for LV blood pool, and -6.7 [-39; 26] g, r = 0.94 for LV wall mass, respectively. Mean relative volume errors of less than 7% were obtained for all chambers. Conclusions: Fully automated assessment of chamber volumes from NCCT scans is feasible and correlates well with volumes obtained from contrast study.

physics.med-ph↗

arXiv2023

Deep Conditional Shape Models for 3D cardiac image segmentation

Delineation of anatomical structures is often the first step of many medical image analysis workflows. While convolutional neural networks achieve high performance, these do not incorporate anatomical shape information. We introduce a novel segmentation algorithm that uses Deep Conditional Shape models (DCSMs) as a core component. Using deep implicit shape representations, the algorithm learns a modality-agnostic shape model that can generate the signed distance functions for any anatomy of interest. To fit the generated shape to the image, the shape model is conditioned on anatomic landmarks that can be automatically detected or provided by the user. Finally, we add a modality-dependent, lightweight refinement network to capture any fine details not represented by the implicit function. The proposed DCSM framework is evaluated on the problem of cardiac left ventricle (LV) segmentation from multiple 3D modalities (contrast-enhanced CT, non-contrasted CT, 3D echocardiography-3DE). We demonstrate that the automatic DCSM outperforms the baseline for non-contrasted CT without the local refinement, and with the refinement for contrasted CT and 3DE, especially with significant improvement in the Hausdorff distance. The semi-automatic DCSM with user-input landmarks, while only trained on contrasted CT, achieves greater than 92% Dice for all modalities. Both automatic DCSM with refinement and semi-automatic DCSM achieve equivalent or better performance compared to inter-user variability for these modalities.

eess.IV↗

arXiv2023

On Equicontinuity and Related Notions in Nonautonomous Dynamical Systems

In this work, we investigate the dynamics of a general non-autonomous system generated by a commutative family of homeomorphisms. In particular, we investigate properties such as periodicity, equicontinuity, minimality and transitivity for a general non-autonomous dynamical system. In \cite{sk2}, the authors derive necessary and sufficient conditions for a system to be minimal. We claim the result to be false and provide an example in support of our claim. Further, we correct the result to derive necessary and sufficient conditions for a non-autonomous system to be minimal. We prove that for an equicontinuous flow generated by a commutative family, while the system need not exhibit almost periodic points, if $x$ is almost periodic then every point in $\overline{\mathcal{O}_H(x)}$ is almost periodic. We further prove that in such a case, the set $\overline{\mathcal{O}_H(x)}$ is uniformly almost periodic and hence provide an analogous extension to a result known for the autonomous systems. We prove that a system generated by a commutative family is transitive if and only if it exhibits a point with dense orbit. We also prove that any minimal system generated by commutative family is either equicontinuous or has a dense set of sensitive points.

math.DS↗

arXiv2023

Automated Cardiac Resting Phase Detection Targeted on the Right Coronary Artery

Static cardiac imaging such as late gadolinium enhancement, mapping, or 3-D coronary angiography require prior information, e.g., the phase during a cardiac cycle with least motion, called resting phase (RP). The purpose of this work is to propose a fully automated framework that allows the detection of the right coronary artery (RCA) RP within CINE series. The proposed prototype system consists of three main steps. First, the localization of the regions of interest (ROI) is performed. Second, the cropped ROI series are taken for tracking motions over all time points. Third, the output motion values are used to classify RPs. In this work, we focused on the detection of the area with the outer edge of the cross-section of the RCA as our target. The proposed framework was evaluated on 102 clinically acquired dataset at 1.5T and 3T. The automatically classified RPs were compared with the reference RPs annotated manually by a expert for testing the robustness and feasibility of the framework. The predicted RCA RPs showed high agreement with the experts annotated RPs with 92.7% accuracy, 90.5% sensitivity and 95.0% specificity for the unseen study dataset. The mean absolute difference of the start and end RP was 13.6 $\pm$ 18.6 ms for the validation study dataset (n=102). In this work, automated RP detection has been introduced by the proposed framework and demonstrated feasibility, robustness, and applicability for static imaging acquisitions.

eess.IV↗

arXiv2022

Sustainable and convenient: bi-modal public transit systems outperforming the private car

Mobility is an indispensable part of modern human societies, but the dominance of motorized individual traffic (MIV, i.e., the private car) leads to a prohibitive waste of energy as well as other resources. Here we show that by combining a line service (e.g., railway) system with a fleet of ride-pooling shuttles connecting line stops to desired pick-up and drop-off points, a bi-modal public transport system may result which provides on-demand door-to-door service at a service level (in terms of transit times) superior to current public transport, and with an overall comfort level comparable to MIV. We identify the conflicting objectives for optimization, i.e., user convenience and energy consumption, and evaluate the system performance in terms of Pareto fronts. By means of simulation and analytical theory, we find that energy consumption can be as low as 20% of MIV, at line service densities typically found in real settings. Surprisingly, we find favorable performance not only in urban, but also in rural settings.

physics.soc-ph↗

arXiv2022

Insights from a pseudospectral study of a potentially singular solution of the three-dimensional axisymmetric incompressible Euler equation

We develop a Fourier-Chebyshev pseudospectral direct numerical simulation (DNS) to examine a potentially singular solution of the radially bounded, three-dimensional (3D), axisymmetric Euler equations [G. Luo and T.Y. Hou, Proc. Natl. Acad. Sci. USA, 111.36 (2014)]. We demonstrate that: (a) the time of singularity is preceded, in any spectrally truncated DNS, by the formation of oscillatory structures called tygers, first investigated in the one-dimensional (1D) Burgers and two-dimensional (2D) Euler equations; (b) the analyticity-strip method can be generalized to obtain an estimate for the (potential) singularity time.

physics.flu-dyn↗

arXiv2021

On Graph Induced Symbolic Systems

\begin{abstract} In this paper, we investigate a shift arising from graph $G$. We prove that any $k$-dimensional shift of finite type can be generated through a $k$-dimensional graph. We investigate the structure of the shift space using the generating matrices for the shift space. We prove that a two dimensional shift space has a horizontally (vertically) periodic point if and only if it possesses a $(m,n)$-periodic point (for some $m,n\in \mathbb{Z}\setminus \{0\}$). We prove that a shift space is finite if and only if it can be generated by permutation matrices. We study the non-emptiness problem and existence of periodic points in terms of the generating matrices.

math.DS↗

arXiv2021

Combo Loss: Handling Input and Output Imbalance in Multi-Organ Segmentation

Simultaneous segmentation of multiple organs from different medical imaging modalities is a crucial task as it can be utilized for computer-aided diagnosis, computer-assisted surgery, and therapy planning. Thanks to the recent advances in deep learning, several deep neural networks for medical image segmentation have been introduced successfully for this purpose. In this paper, we focus on learning a deep multi-organ segmentation network that labels voxels. In particular, we examine the critical choice of a loss function in order to handle the notorious imbalance problem that plagues both the input and output of a learning model. The input imbalance refers to the class-imbalance in the input training samples (i.e., small foreground objects embedded in an abundance of background voxels, as well as organs of varying sizes). The output imbalance refers to the imbalance between the false positives and false negatives of the inference model. In order to tackle both types of imbalance during training and inference, we introduce a new curriculum learning based loss function. Specifically, we leverage Dice similarity coefficient to deter model parameters from being held at bad local minima and at the same time gradually learn better model parameters by penalizing for false positives/negatives using a cross entropy term. We evaluated the proposed loss function on three datasets: whole body positron emission tomography (PET) scans with 5 target organs, magnetic resonance imaging (MRI) prostate scans, and ultrasound echocardigraphy images with a single target organ i.e., left ventricular. We show that a simple network architecture with the proposed integrative loss function can outperform state-of-the-art methods and results of the competing methods can be improved when our proposed loss is used.

cs.CV↗

arXiv2020

Spatial Sharing of GPU for Autotuning DNN models

GPUs are used for training, inference, and tuning the machine learning models. However, Deep Neural Network (DNN) vary widely in their ability to exploit the full power of high-performance GPUs. Spatial sharing of GPU enables multiplexing several DNNs on the GPU and can improve GPU utilization, thus improving throughput and lowering latency. DNN models given just the right amount of GPU resources can still provide low inference latency, just as much as dedicating all of the GPU for their inference task. An approach to improve DNN inference is tuning of the DNN model. Autotuning frameworks find the optimal low-level implementation for a certain target device based on the trained machine learning model, thus reducing the DNN's inference latency and increasing inference throughput. We observe an interdependency between the tuned model and its inference latency. A DNN model tuned with specific GPU resources provides the best inference latency when inferred with close to the same amount of GPU resources. While a model tuned with the maximum amount of the GPU's resources has poorer inference latency once the GPU resources are limited for inference. On the other hand, a model tuned with an appropriate amount of GPU resources still achieves good inference latency across a wide range of GPU resource availability. We explore the causes that impact the tuning of a model at different amounts of GPU resources. We present many techniques to maximize resource utilization and improve tuning performance. We enable controlled spatial sharing of GPU to multiplex several tuning applications on the GPU. We scale the tuning server instances and shard the tuning model across multiple client instances for concurrent tuning of different operators of a model, achieving better GPU multiplexing. With our improvements, we decrease DNN autotuning time by up to 75 percent and increase throughput by a factor of 5.

cs.NE↗