Review of Artificial Intelligence Techniques in Imaging Data Acquisition, Segmentation, and Diagnosis for COVID-19

IEEE Reviews in Biomedical Engineering, Journal Year: 2020, Volume and Issue: 14, P. 4 - 15

Published: April 16, 2020

The pandemic of coronavirus disease 2019 (COVID-19) is spreading all over the world. Medical imaging such as X-ray and computed tomography (CT) plays an essential role in global fight against COVID-19, whereas recently emerging artificial intelligence (AI) technologies further strengthen power tools help medical specialists. We hereby review rapid responses community (empowered by AI) toward COVID-19. For example, AI-empowered image acquisition can significantly automate scanning procedure also reshape workflow with minimal contact to patients, providing best protection technicians. Also, AI improve work efficiency accurate delineation infections CT images, facilitating subsequent quantification. Moreover, computer-aided platforms radiologists make clinical decisions, i.e., for diagnosis, tracking, prognosis. In this paper, we thus cover entire pipeline analysis techniques involved including acquisition, segmentation, follow-up. particularly focus on integration CT, both which are widely used frontline hospitals, order depict latest progress radiology fighting

Language: Английский

---

Inf-Net: Automatic COVID-19 Lung Infection Segmentation From CT Images

IEEE Transactions on Medical Imaging, Journal Year: 2020, Volume and Issue: 39(8), P. 2626 - 2637

Published: May 22, 2020

Coronavirus Disease 2019 (COVID-19) spread globally in early 2020, causing the world to face an existential health crisis. Automated detection of lung infections from computed tomography (CT) images offers a great potential augment traditional healthcare strategy for tackling COVID-19. However, segmenting infected regions CT slices faces several challenges, including high variation infection characteristics, and low intensity contrast between normal tissues. Further, collecting large amount data is impractical within short time period, inhibiting training deep model. To address these novel COVID-19 Lung Infection Segmentation Deep Network ( Inf-Net ) proposed automatically identify chest slices. In our , parallel partial decoder used aggregate high-level features generate global map. Then, implicit reverse attention explicit edge-attention are utilized model boundaries enhance representations. Moreover, alleviate shortage labeled data, we present semi-supervised segmentation framework based on randomly selected propagation strategy, which only requires few leverages primarily unlabeled data. Our can improve learning ability achieve higher performance. Extensive experiments xmlns:xlink="http://www.w3.org/1999/xlink">COVID-SemiSeg real volumes demonstrate that outperforms most cutting-edge models advances state-of-the-art

Language: Английский

---

Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans

Nature Machine Intelligence, Journal Year: 2021, Volume and Issue: 3(3), P. 199 - 217

Published: March 15, 2021

Machine learning methods offer great promise for fast and accurate detection prognostication of COVID-19 from standard-of-care chest radiographs (CXR) computed tomography (CT) images. Many articles have been published in 2020 describing new machine learning-based models both these tasks, but it is unclear which are potential clinical utility. In this systematic review, we search EMBASE via OVID, MEDLINE PubMed, bioRxiv, medRxiv arXiv papers preprints uploaded January 1, to October 3, describe the diagnosis or prognosis CXR CT Our identified 2,212 studies, 415 were included after initial screening and, quality screening, 61 studies review. review finds that none use due methodological flaws and/or underlying biases. This a major weakness, given urgency with validated needed. To address this, give many recommendations which, if followed, will solve issues lead higher model development well documented manuscripts.

Language: Английский

---

Multi-task deep learning based CT imaging analysis for COVID-19 pneumonia: Classification and segmentation

Computers in Biology and Medicine, Journal Year: 2020, Volume and Issue: 126, P. 104037 - 104037

Published: Oct. 8, 2020

Language: Английский

---

COVID-19 Detection Through Transfer Learning Using Multimodal Imaging Data

IEEE Access, Journal Year: 2020, Volume and Issue: 8, P. 149808 - 149824

Published: Jan. 1, 2020

Detecting COVID-19 early may help in devising an appropriate treatment plan and disease containment decisions. In this study, we demonstrate how transfer learning from deep models can be used to perform detection using images three most commonly medical imaging modes X-Ray, Ultrasound, CT scan. The aim is provide over-stressed professionals a second pair of eyes through intelligent image classification models. We identify suitable Convolutional Neural Network (CNN) model initial comparative study several popular CNN then optimize the selected VGG19 for modalities show highly scarce challenging datasets. highlight challenges (including dataset size quality) utilizing current publicly available datasets developing useful it adversely impacts trainability complex also propose pre-processing stage create trustworthy testing new approach aimed reduce unwanted noise so that focus on detecting diseases with specific features them. Our results indicate Ultrasound superior accuracy compared X-Ray scans. experimental limited data, deeper networks struggle train well provides less consistency over are using. model, which extensively tuned parameters, performs considerable levels against pneumonia or normal all lung precision up 86% 100% 84%

Language: Английский

---

Sample-Efficient Deep Learning for COVID-19 Diagnosis Based on CT Scans

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2020, Volume and Issue: unknown

Published: April 17, 2020

Abstract Coronavirus disease 2019 (COVID-19) has infected more than 1.3 million individuals all over the world and caused 106,000 deaths. One major hurdle in controlling spreading of this is inefficiency shortage medical tests. There have been increasing efforts on developing deep learning methods to diagnose COVID-19 based CT scans. However, these works are difficult reproduce adopt since data used their studies not publicly available. Besides, require a large number CTs train accurate diagnosis models, which obtain. In paper, we aim address two problems. We build publicly-available dataset containing hundreds scans positive for develop sample-efficient that can achieve high accuracy from even when training images limited. Specifically, propose Self-Trans approach, synergistically integrates contrastive self-supervised with transfer learn powerful unbiased feature representations reducing risk overfitting. Extensive experiments demonstrate superior performance our proposed approach compared several state-of-the-art baselines. Our achieves an F1 0.85 AUC 0.94 diagnosing scans, though just few hundred.

Language: Английский

---

Blockchain-Federated-Learning and Deep Learning Models for COVID-19 Detection Using CT Imaging

IEEE Sensors Journal, Journal Year: 2021, Volume and Issue: 21(14), P. 16301 - 16314

Published: April 30, 2021

With the increase of COVID-19 cases worldwide, an effective way is required to diagnose patients. The primary problem in diagnosing patients shortage and reliability testing kits, due quick spread virus, medical practitioners are facing difficulty identifying positive cases. second real-world share data among hospitals globally while keeping view privacy concerns organizations. Building a collaborative model preserving major for training global deep learning model. This paper proposes framework that collects small amount from different sources (various hospitals) trains using blockchain based federated learning. Blockchain technology authenticates organization. First, we propose normalization technique deals with heterogeneity as gathered having kinds CT scanners. Secondly, use Capsule Network-based segmentation classification detect Thirdly, design method can collaboratively train privacy. Additionally, collected real-life data, which is, open research community. proposed utilize up-to-date improves recognition computed tomography (CT) images. Finally, our results demonstrate better performance

Language: Английский

---

Iteratively Pruned Deep Learning Ensembles for COVID-19 Detection in Chest X-Rays

IEEE Access, Journal Year: 2020, Volume and Issue: 8, P. 115041 - 115050

Published: Jan. 1, 2020

We demonstrate use of iteratively pruned deep learning model ensembles for detecting pulmonary manifestation COVID-19 with chest X-rays. This disease is caused by the novel Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) virus, also known as (2019-nCoV). A custom convolutional neural network and a selection ImageNet pretrained models are trained evaluated at patient-level on publicly available CXR collections to learn modality-specific feature representations. The learned knowledge transferred fine-tuned improve performance generalization in related task classifying CXRs normal, showing bacterial pneumonia, or COVID-19-viral abnormalities. best performing reduce complexity memory efficiency. predictions best-performing combined through different ensemble strategies classification performance. Empirical evaluations that weighted average significantly improves resulting an accuracy 99.01% area under curve 0.9972 findings CXRs. transfer, iterative pruning, resulted improved predictions. expect this can be quickly adopted screening using radiographs.

Language: Английский

---

AI-assisted CT imaging analysis for COVID-19 screening: Building and deploying a medical AI system

Applied Soft Computing, Journal Year: 2020, Volume and Issue: 98, P. 106897 - 106897

Published: Nov. 10, 2020

Language: Английский

---

A review of mathematical modeling, artificial intelligence and datasets used in the study, prediction and management of COVID-19

Applied Intelligence, Journal Year: 2020, Volume and Issue: 50(11), P. 3913 - 3925

Published: July 6, 2020

In the past few months, several works were published in regards to dynamics and early detection of COVID-19 via mathematical modeling Artificial intelligence (AI). The aim this work is provide research community with comprehensive overview methods used these studies as well a compendium available open source datasets COVID-19. all, 61 journal articles, reports, fact sheets, websites dealing studied reviewed. It was found that most done based on Susceptible-Exposed-Infected-Removed (SEIR) Susceptible-infected-recovered (SIR) models while AI implementations Convolutional Neural Network (CNN) X-ray CT images. terms datasets, they include aggregated case medical images, management strategies, healthcare workforce, demography, mobility during outbreak. Both Mathematical have both shown be reliable tools fight against pandemic. Several concerning also been collected shared source. However, much needed diversification datasets. Other applications should explored

Language: Английский

---