Review of deep learning: concepts, CNN architectures, challenges, applications, future directions

Journal Of Big Data, Journal Year: 2021, Volume and Issue: 8(1)

Published: March 31, 2021

In the last few years, deep learning (DL) computing paradigm has been deemed Gold Standard in machine (ML) community. Moreover, it gradually become most widely used computational approach field of ML, thus achieving outstanding results on several complex cognitive tasks, matching or even beating those provided by human performance. One benefits DL is ability to learn massive amounts data. The grown fast years and extensively successfully address a wide range traditional applications. More importantly, outperformed well-known ML techniques many domains, e.g., cybersecurity, natural language processing, bioinformatics, robotics control, medical information among others. Despite contributed works reviewing State-of-the-Art DL, all them only tackled one aspect which leads an overall lack knowledge about it. Therefore, this contribution, we propose using more holistic order provide suitable starting point from develop full understanding DL. Specifically, review attempts comprehensive survey important aspects including enhancements recently added field. particular, paper outlines importance presents types networks. It then convolutional neural networks (CNNs) utilized network type describes development CNNs architectures together with their main features, AlexNet closing High-Resolution (HR.Net). Finally, further present challenges suggested solutions help researchers understand existing research gaps. followed list major Computational tools FPGA, GPU, CPU are summarized along description influence ends evolution matrix, benchmark datasets, summary conclusion.

Language: Английский

---

COVID-Net: a tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images

Scientific Reports, Journal Year: 2020, Volume and Issue: 10(1)

Published: Nov. 11, 2020

Abstract The Coronavirus Disease 2019 (COVID-19) pandemic continues to have a devastating effect on the health and well-being of global population. A critical step in fight against COVID-19 is effective screening infected patients, with one key approaches being radiology examination using chest radiography. It was found early studies that patients present abnormalities radiography images are characteristic those COVID-19. Motivated by this inspired open source efforts research community, study we introduce COVID-Net, deep convolutional neural network design tailored for detection cases from X-ray (CXR) available general public. To best authors’ knowledge, COVID-Net first designs CXR at time initial release. We also COVIDx, an access benchmark dataset generated comprising 13,975 across 13,870 patient cases, largest number publicly positive knowledge. Furthermore, investigate how makes predictions explainability method attempt not only gain deeper insights into factors associated COVID which can aid clinicians improved screening, but audit responsible transparent manner validate it making decisions based relevant information images. By no means production-ready solution, hope along description constructing COVIDx dataset, will be leveraged build upon both researchers citizen data scientists alike accelerate development highly accurate yet practical learning solutions detecting treatment who need most.

Language: Английский

---

Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal

BMJ, Journal Year: 2020, Volume and Issue: unknown, P. m1328 - m1328

Published: April 7, 2020

To review and appraise the validity usefulness of published preprint reports prediction models for diagnosing coronavirus disease 2019 (covid-19) in patients with suspected infection, prognosis covid-19, detecting people general population at increased risk covid-19 infection or being admitted to hospital disease.

Language: Английский

---

Automated detection of COVID-19 cases using deep neural networks with X-ray images

Computers in Biology and Medicine, Journal Year: 2020, Volume and Issue: 121, P. 103792 - 103792

Published: April 28, 2020

Language: Английский

---

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

---

Automatic detection of coronavirus disease (COVID-19) using X-ray images and deep convolutional neural networks

Pattern Analysis and Applications, Journal Year: 2021, Volume and Issue: 24(3), P. 1207 - 1220

Published: May 9, 2021

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

---

COVIDiagnosis-Net: Deep Bayes-SqueezeNet based diagnosis of the coronavirus disease 2019 (COVID-19) from X-ray images

Medical Hypotheses, Journal Year: 2020, Volume and Issue: 140, P. 109761 - 109761

Published: April 23, 2020

Language: Английский

---

COVID-CAPS: A capsule network-based framework for identification of COVID-19 cases from X-ray images

Pattern Recognition Letters, Journal Year: 2020, Volume and Issue: 138, P. 638 - 643

Published: Sept. 16, 2020

Language: Английский

---

COVID-CT-Dataset: A CT Scan Dataset about COVID-19

arXiv (Cornell University), Journal Year: 2020, Volume and Issue: unknown

Published: Jan. 1, 2020

During the outbreak time of COVID-19, computed tomography (CT) is a useful manner for diagnosing COVID-19 patients. Due to privacy issues, publicly available CT datasets are highly difficult obtain, which hinders research and development AI-powered diagnosis methods based on CTs. To address this issue, we build an open-sourced dataset -- COVID-CT, contains 349 images from 216 patients 463 non-COVID-19 The utility confirmed by senior radiologist who has been treating since pandemic. We also perform experimental studies further demonstrate that developing AI-based models COVID-19. Using dataset, develop multi-task learning self-supervised learning, achieve F1 0.90, AUC 0.98, accuracy 0.89. According radiologist, with such performance good enough clinical usage. data code at https://github.com/UCSD-AI4H/COVID-CT

Language: Английский

---