About: Interpreting Deep Ensemble Learning through Radiologist Annotations for COVID-19 Detection in Chest Radiographs

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Interpreting Deep Ensemble Learning through Radiologist Annotations for COVID-19 Detection in Chest Radiographs Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
title	Interpreting Deep Ensemble Learning through Radiologist Annotations for COVID-19 Detection in Chest Radiographs
Creator	Antani, Sameer Rajaraman, Sivaramakrishnan Alderson, Philip Folio, Les Sornapudi, Sudhir
source	MedRxiv
abstract	Data-driven deep learning (DL) methods using convolutional neural networks (CNNs) demonstrate promising performance in natural image computer vision tasks. However, using these models in medical computer vision tasks suffers from several limitations, viz., (i) adapting to visual characteristics that are unlike natural images; (ii) modeling random noise during training due to stochastic optimization and backpropagation-based learning strategy; (iii) challenges in explaining DL black-box behavior to support clinical decision-making; and (iv) inter-reader variability in the ground truth (GT) annotations affecting learning and evaluation. This study proposes a systematic approach to address these limitations for COVID-19 detection using chest X-rays (CXRs). Specifically, our contribution benefits from (i) pretraining specific to CXRs in transferring and fine-tuning the learned knowledge toward improving COVID-19 detection performance; (ii) using ensembles of the fine-tuned models to further improve performance compared to individual constituent models; (iii) performing statistical analyses at various learning stages to validate our claims; (iv) interpreting learned individual and ensemble model behavior through class-selective relevance mapping (CRM)-based region of interest (ROI) localization; (v) analyzing inter-reader variability and ensemble localization performance using Simultaneous Truth and Performance Level Estimation (STAPLE) methods. We observe that: (i) ensemble approaches improved classification and localization performance; and, (ii) inter-reader variability and performance level assessment helped guide algorithm design and parameter optimization. To the best of our knowledge, this is the first study to construct ensembles, perform ensemble-based disease ROI localization, and analyze inter-reader variability and algorithm performance for COVID-19 detection in CXRs.
has issue date	2020-07-16(xsd:dateTime)
bibo:doi	10.1101/2020.07.15.20154385
has license	medrxiv
sha1sum (hex)	c2d15ed4da2dc2d3d546c29a469565d84528958f
schema:url	https://doi.org/10.1101/2020.07.15.20154385
resource representing a document's title	Interpreting Deep Ensemble Learning through Radiologist Annotations for COVID-19 Detection in Chest Radiographs
resource representing a document's body	covid:c2d15ed4da2dc2d3d546c29a469565d84528958f#body_text
is schema:about of	named entity 'guide' named entity 'promising' named entity 'ensembles' named entity 'Simultaneous' named entity 'demonstrate' named entity 'stochastic optimization' named entity 'model behavior' named entity 'training' named entity 'detection' named entity 'learning' named entity 'COVID-19' named entity 'annotations' named entity 'transferring' named entity 'variability' named entity 'support' named entity 'specific' named entity 'images' named entity 'COVID-19' named entity 'disease' named entity 'localization' named entity 'stochastic optimization' named entity 'random noise' named entity 'ensemble model' named entity 'clinical decision-making' named entity 'CRM' named entity 'performance' named entity 'ImageNet' named entity 'WRN' named entity 'CNNs' named entity 'CRM' named entity 'COVID-19' named entity 'ImageNet' named entity 'classifiers' named entity 'causality' named entity 'CUDA' named entity 'U-Net' named entity 'COVID-19' named entity 'residual network' named entity 'viral pneumonia' named entity 'training data' named entity 'radiologist' named entity 'ensemble learning' named entity 'CNNs' named entity 'regularization' named entity 'statistical analyses' named entity 'COVID-19' named entity 'IoU' named entity 'neural networks' named entity 'COVID-19' named entity 'Area under curve' named entity 'algorithm' named entity 'CNN' named entity 'viral pneumonia' named entity 'COVID-19' named entity 'ImageNet' named entity 'viral pneumonia' named entity 'WRN' named entity 'ensemble learning' named entity 'algorithm' named entity 'British Journal of Pharmacology' named entity 'Ensemble models' named entity 'decision-making' named entity 'transfer learning' named entity 'COVID-19' named entity 'F-distribution' named entity 'CT scans' named entity 'Nanchong' named entity 'learned behavior' named entity 'AUC values' named entity 'MCC' named entity '95% CI' named entity 'radiologists' named entity 'one-way ANOVA' named entity 'medical images' named entity 'IoU' named entity 'statistical significance' named entity 'COVID-19' named entity 'statistical analysis' named entity 'inter-observer variability' named entity 'ROIs' named entity 'radiologists'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2025 OpenLink Software