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

---

Recent advances and clinical applications of deep learning in medical image analysis

Medical Image Analysis, Journal Year: 2022, Volume and Issue: 79, P. 102444 - 102444

Published: April 4, 2022

Language: Английский

---

Secure and Provenance Enhanced Internet of Health Things Framework: A Blockchain Managed Federated Learning Approach

IEEE Access, Journal Year: 2020, Volume and Issue: 8, P. 205071 - 205087

Published: Jan. 1, 2020

Recent advancements in the Internet of Health Things (IoHT) have ushered wide adoption IoT devices our daily health management. For IoHT data to be acceptable by stakeholders, applications that incorporate must a provision for provenance, addition accuracy, security, integrity, and quality data. To protect privacy security data, federated learning (FL) differential (DP) been proposed, where private can trained at owner’s premises. hardware GPUs even allow FL process within smartphone or edge having attached their nodes. Although some concerns are addressed FL, fully decentralized is still challenge due lack training capability all nodes, scarcity high-quality datasets, provenance authentication required each node. In this paper, we present lightweight hybrid framework which blockchain smart contracts manage plan, trust management, participating distribution global locally models, reputation nodes uploaded datasets models. The also supports full encryption dataset, model training, inferencing process. Each node performs additive encryption, while uses multiplicative aggregate updated parameters. support anonymization DP. This was tested with several deep designed clinical trials COVID-19 patients. We here detailed design, implementation, test results, demonstrate strong potential wider IoHT-based management secure way.

Language: Английский

---

Medical image-based detection of COVID-19 using Deep Convolution Neural Networks

Multimedia Systems, Journal Year: 2021, Volume and Issue: 29(3), P. 1729 - 1738

Published: April 28, 2021

The demand for automatic detection of Novel Coronavirus or COVID-19 is increasing across the globe. exponential rise in cases burdens healthcare facilities, and a vast amount multimedia data being explored to find solution. This study presents practical solution detect from chest X-rays while distinguishing those normal impacted by Viral Pneumonia via Deep Convolution Neural Networks (CNN). In this study, three pre-trained CNN models (EfficientNetB0, VGG16, InceptionV3) are evaluated through transfer learning. rationale selecting these specific their balance accuracy efficiency with fewer parameters suitable mobile applications. dataset used publicly available compiled different sources. uses deep learning techniques performance metrics (accuracy, recall, specificity, precision, F1 scores). results show that proposed approach produced high-quality model, an overall 92.93%, COVID-19, sensitivity 94.79%. work indicates definite possibility implement computer vision design enable effective screening measures.

Language: Английский

---

An Internet-of-Medical-Things-Enabled Edge Computing Framework for Tackling COVID-19

IEEE Internet of Things Journal, Journal Year: 2021, Volume and Issue: 8(21), P. 15847 - 15854

Published: Jan. 13, 2021

Capturing psychological, emotional, and physiological states, especially during a pandemic, leveraging the captured sensory data within pandemic management ecosystem is challenging. Recent advancements for Internet of Medical Things (IoMT) have shown promising results from collecting diversified types such emotional physical health-related home environment. State-of-the-art deep learning (DL) applications can run in resource-constrained edge environment, which allows IoMT devices to be processed locally at edge, performs inferencing related in-home health. This health remain vicinity user while ensuring privacy, security, low latency system. In this article, we develop an system that uses DL detect COVID-19 symptoms generates reports alerts used medical decision support. Several been developed, tested, deployed support clinical trials. We present design framework, description our implemented system, accuracy results. The test show suitability pandemic.

Language: Английский

---

MetaMed: Few-shot medical image classification using gradient-based meta-learning

Pattern Recognition, Journal Year: 2021, Volume and Issue: 120, P. 108111 - 108111

Published: June 17, 2021

Language: Английский

---

COVID-19 and Non-COVID-19 Classification using Multi-layers Fusion From Lung Ultrasound Images

Information Fusion, Journal Year: 2021, Volume and Issue: 72, P. 80 - 88

Published: Feb. 24, 2021

Language: Английский

---

Meta-learning approaches for learning-to-learn in deep learning: A survey

Neurocomputing, Journal Year: 2022, Volume and Issue: 494, P. 203 - 223

Published: April 21, 2022

Language: Английский

---

Federated learning for secure IoMT-applications in smart healthcare systems: A comprehensive review

Knowledge-Based Systems, Journal Year: 2023, Volume and Issue: 274, P. 110658 - 110658

Published: May 22, 2023

Recent developments in the Internet of Things (IoT) and various communication technologies have reshaped numerous application areas. Nowadays, IoT is assimilated into medical devices equipment, leading to progression Medical (IoMT). Therefore, IoMT-based healthcare applications are deployed used day-to-day scenario. Traditionally, machine learning (ML) models use centralized data compilation that impractical pragmatic frameworks due rising privacy security issues. Federated Learning (FL) has been observed as a developing distributed collective paradigm, most appropriate for modern framework, manages stakeholders (e.g., patients, hospitals, laboratories, etc.) carry out training without actual exchange sensitive data. Consequently, this work, authors present an exhaustive survey on FL-based IoMT smart frameworks. First, introduced devices, their types, applications, datasets, framework detail. Subsequently, concept FL, its domains, tools develop FL discussed. The significant contribution deploying secure systems presented by focusing patents, real-world projects, datasets. A comparison techniques with other schemes ecosystem also presented. Finally, discussed challenges faced potential future research recommendations deploy

Language: Английский

---

Four-way classification of Alzheimer’s disease using deep Siamese convolutional neural network with triplet-loss function

Brain Informatics, Journal Year: 2023, Volume and Issue: 10(1)

Published: Feb. 17, 2023

Alzheimer's disease (AD) is a neurodegenerative that causes irreversible damage to several brain regions, including the hippocampus causing impairment in cognition, function, and behaviour. Early diagnosis of will reduce suffering patients their family members. Towards this aim, paper, we propose Siamese Convolutional Neural Network (SCNN) architecture employs triplet-loss function for representation input MRI images as k-dimensional embeddings. We used both pre-trained non-pretrained CNNs transform into embedding space. These embeddings are subsequently 4-way classification disease. The model efficacy was tested using ADNI OASIS datasets which produced an accuracy 91.83% 93.85%, respectively. Furthermore, obtained results compared with similar methods proposed literature.

Language: Английский

---