arXiv subjects

Arjun Sharma

Publications and source records attributed to Arjun Sharma.

27 records · Page 2Linked to original sources

arXiv2020

Chest X-ray Report Generation through Fine-Grained Label Learning

Obtaining automated preliminary read reports for common exams such as chest X-rays will expedite clinical workflows and improve operational efficiencies in hospitals. However, the quality of reports generated by current automated approaches is not yet clinically acceptable as they cannot ensure the correct detection of a broad spectrum of radiographic findings nor describe them accurately in terms of laterality, anatomical location, severity, etc. In this work, we present a domain-aware automatic chest X-ray radiology report generation algorithm that learns fine-grained description of findings from images and uses their pattern of occurrences to retrieve and customize similar reports from a large report database. We also develop an automatic labeling algorithm for assigning such descriptors to images and build a novel deep learning network that recognizes both coarse and fine-grained descriptions of findings. The resulting report generation algorithm significantly outperforms the state of the art using established score metrics.

cs.CV↗

arXiv2020

Early evolution of optimal perturbations in a viscosity-stratified channel

This work shows how the early stages of perturbation growth in a viscosity-stratified flow are different from those in a constant-viscosity flow, and how nonlinearity is a crucial ingredient. We derive the viscosity-varying adjoint Navier-Stokes equations, where gradients in viscosity force both the adjoint momentum and the adjoint scalar (here temperature). By the technique of direct-adjoint looping, we obtain the nonlinear optimal perturbation which maximises the perturbation kinetic energy of the nonlinear system. While we study three-dimensional plane Poiseuille (channel) flow with the walls at different temperatures, and a temperature-dependent viscosity, our findings are general for any flow with viscosity variations near walls. The Orr and modified lift-up mechanisms are in operation at low and high perturbation amplitudes respectively at our subcritical Reynolds number. The nonlinear optimal perturbation contains more energy on the hot (less-viscous) side, with a stronger initial lift-up. However, as the flow evolves, the important dynamics shifts to the cold (more-viscous) side, where wide high-speed streaks of low viscosity grow and persist, and strengthen the inflectional quality of the velocity profile. We provide a physical description of this process, and show that the evolution of the linear optimal perturbation misses most of the physics. The Prandtl number does not qualitatively affect the findings at these times. The study of nonlinear optimal perturbations is still in its infancy, and viscosity variations are ubiquitous. We hope that this first work on nonlinear optimal perturbation with viscosity variations will lead to wider studies on transition to turbulence in these flows.

physics.flu-dyn↗

arXiv2020

A robust network architecture to detect normal chest X-ray radiographs

We propose a novel deep neural network architecture for normalcy detection in chest X-ray images. This architecture treats the problem as fine-grained binary classification in which the normal cases are well-defined as a class while leaving all other cases in the broad class of abnormal. It employs several components that allow generalization and prevent overfitting across demographics. The model is trained and validated on a large public dataset of frontal chest X-ray images. It is then tested independently on images from a clinical institution of differing patient demographics using a three radiologist consensus for ground truth labeling. The model provides an area under ROC curve of 0.96 when tested on 1271 images. We can automatically remove nearly a third of disease-free chest X-ray screening images from the workflow, without introducing any false negatives (100% sensitivity to disease) thus raising the potential of expediting radiology workflows in hospitals in future.

eess.IV↗

arXiv2019

Automated Detection and Type Classification of Central Venous Catheters in Chest X-Rays

Central venous catheters (CVCs) are commonly used in critical care settings for monitoring body functions and administering medications. They are often described in radiology reports by referring to their presence, identity and placement. In this paper, we address the problem of automatic detection of their presence and identity through automated segmentation using deep learning networks and classification based on their intersection with previously learned shape priors from clinician annotations of CVCs. The results not only outperform existing methods of catheter detection achieving 85.2% accuracy at 91.6% precision, but also enable high precision (95.2%) classification of catheter types on a large dataset of over 10,000 chest X-rays, presenting a robust and practical solution to this problem.

eess.IV↗

arXiv2019

Analytic Reconstruction of a Two-Dimensional Velocity Field from an Observed Diffusive Scalar

Inverting an evolving diffusive scalar field to reconstruct the underlying velocity field is an underdetermined problem. Here we show, however, that for two-dimensional incompressible flows, this inverse problem can still be uniquely solved if high-resolution tracer measurements, as well as velocity measurements along a curve transverse to the instantaneous scalar contours, are available. Such measurements enable solving a system of partial differential equations for the velocity components by the method of characteristics. If the value of the scalar diffusivity is known, then knowledge of just one velocity component along a transverse initial curve is sufficient. These conclusions extend to the shallow-water equations and to flows with spatially dependent diffusivity. We illustrate our results on velocity reconstruction from tracer fields for planar Navier- Stokes flows and for a barotropic ocean circulation model. We also discuss the use of the proposed velocity reconstruction in oceanographic applications to extend localised velocity measurements to larger spatial domains with the help of remotely sensed scalar fields.

physics.flu-dyn↗

arXiv2019

Directed-Info GAIL: Learning Hierarchical Policies from Unsegmented Demonstrations using Directed Information

The use of imitation learning to learn a single policy for a complex task that has multiple modes or hierarchical structure can be challenging. In fact, previous work has shown that when the modes are known, learning separate policies for each mode or sub-task can greatly improve the performance of imitation learning. In this work, we discover the interaction between sub-tasks from their resulting state-action trajectory sequences using a directed graphical model. We propose a new algorithm based on the generative adversarial imitation learning framework which automatically learns sub-task policies from unsegmented demonstrations. Our approach maximizes the directed information flow in the graphical model between sub-task latent variables and their generated trajectories. We also show how our approach connects with the existing Options framework, which is commonly used to learn hierarchical policies.

cs.LG↗

arXiv2019

Age prediction using a large chest X-ray dataset

Age prediction based on appearances of different anatomies in medical images has been clinically explored for many decades. In this paper, we used deep learning to predict a persons age on Chest X-Rays. Specifically, we trained a CNN in regression fashion on a large publicly available dataset. Moreover, for interpretability, we explored activation maps to identify which areas of a CXR image are important for the machine (i.e. CNN) to predict a patients age, offering insight. Overall, amongst correctly predicted CXRs, we see areas near the clavicles, shoulders, spine, and mediastinum being most activated for age prediction, as one would expect biologically. Amongst incorrectly predicted CXRs, we have qualitatively identified disease patterns that could possibly make the anatomies appear older or younger than expected. A further technical and clinical evaluation would improve this work. As CXR is the most commonly requested imaging exam, a potential use case for estimating age may be found in the preventative counseling of patient health status compared to their age-expected average, particularly when there is a large discrepancy between predicted age and the real patient age.

cs.CV↗

arXiv2016

GeThR-Net: A Generalized Temporally Hybrid Recurrent Neural Network for Multimodal Information Fusion

Data generated from real world events are usually temporal and contain multimodal information such as audio, visual, depth, sensor etc. which are required to be intelligently combined for classification tasks. In this paper, we propose a novel generalized deep neural network architecture where temporal streams from multiple modalities are combined. There are total M+1 (M is the number of modalities) components in the proposed network. The first component is a novel temporally hybrid Recurrent Neural Network (RNN) that exploits the complimentary nature of the multimodal temporal information by allowing the network to learn both modality specific temporal dynamics as well as the dynamics in a multimodal feature space. M additional components are added to the network which extract discriminative but non-temporal cues from each modality. Finally, the predictions from all of these components are linearly combined using a set of automatically learned weights. We perform exhaustive experiments on three different datasets spanning four modalities. The proposed network is relatively 3.5%, 5.7% and 2% better than the best performing temporal multimodal baseline for UCF-101, CCV and Multimodal Gesture datasets respectively.

cs.CV↗

arXiv2013

Viability of detection by AMS of sudden features due to dark matter annihilation to positrons and electrons

The Fermi experiment has measured the cosmic ray electron+positron spectrum and positron fraction [$Φ_{e^+}/(Φ_{e^+ + e^-})]$, and PAMELA has measured the positron fraction with better precision. While the majority of cosmic ray electrons and positrons are of astrophysical origin, there may also be a contribution from dark matter annihilation in the galactic halo. The upcoming results of the AMS experiment will show measurements of these quantities with far greater precision. One dark matter annihilation scenario is where two dark matter particles annihilate directly to $e^+$ and $e^-$ final states. In this article, we calculate the signature "bumps" in these measurements assuming a given density profile (NFW profile). If the dark matter annihilates to electrons and positrons with a cross section $σv \sim\, 10^{-26}$ cm$^3$/s or greater, this feature may be discernible by AMS. However, we demonstrate that such a prominent spectral feature is already ruled out by the relative smoothness of the positron + electron cosmic ray spectrum as measured by Fermi. Hence we conclude that such a feature is undetectable unless the mass is less than $\sim$40 GeV.

astro-ph.CO↗