Showing papers in "IEEE Journal of Biomedical and Health Informatics in 2022"
TL;DR: Wang et al. as mentioned in this paper proposed an efficient and large-scale batch verification scheme with group testing technology based on ECDSA, which uses group testing to improve the efficiency of identifying invalid signatures.
Abstract: The blockchain-enabled internet of medical things (IoMT) is an emerging paradigm that could provide strong trust establishment and ensure the traceability of data sharing in the IoMT networks. One of the fundamental building blocks for Blockchain is Elliptic Curve Digital Signature Algorithm (ECDSA). Nevertheless, when processing a large number of transactions, the verification of multiple signatures will incur cumbersome overhead to the nodes in Blockchain. Although batch verification is able to provide a promising approach that verifies multiple signatures simultaneously and efficiently, the upper bound of batch size is limited to small-scale and the efficiency will drop rapidly as the batch size grows in the state-of-the-art ECDSA batch schemes. Meanwhile, most of the existing researches only focus on improving the efficiency of batch verification algorithms in various cryptosystem while ignoring the identification of invalid signatures, which could cause severe performance degradation when the batch verification fails. Motivated by these observations, this paper proposes an efficient and large-scale batch verification scheme with group testing technology based on ECDSA. The application of the presented protocols in Bitcoin and Hyperledger Fabric has been analyzed as supportive and effective. When the batch verification returns a false result, we utilize group testing technology to improve the efficiency of identifying invalid signatures. Comprehensive simulation results demonstrate that our protocol outperforms the related ECDSA batch verification schemes.
59 citations
TL;DR: Wang et al. as discussed by the authors proposed a multi-scale residual fusion network (MSRF-Net) for medical image segmentation, which is able to exchange multiscale features of varying receptive fields using a Dual-Scale Dense Fusion (DSDF) block.
Abstract: Methods based on convolutional neural networks have improved the performance of biomedical image segmentation. However, most of these methods cannot efficiently segment objects of variable sizes and train on small and biased datasets, which are common for biomedical use cases. While methods exist that incorporate multi-scale fusion approaches to address the challenges arising with variable sizes, they usually use complex models that are more suitable for general semantic segmentation problems. In this paper, we propose a novel architecture called Multi-Scale Residual Fusion Network (MSRF-Net), which is specially designed for medical image segmentation. The proposed MSRF-Net is able to exchange multi-scale features of varying receptive fields using a Dual-Scale Dense Fusion (DSDF) block. Our DSDF block can exchange information rigorously across two different resolution scales, and our MSRF sub-network uses multiple DSDF blocks in sequence to perform multi-scale fusion. This allows the preservation of resolution, improved information flow and propagation of both high- and low-level features to obtain accurate segmentation maps. The proposed MSRF-Net allows to capture object variabilities and provides improved results on different biomedical datasets. Extensive experiments on MSRF-Net demonstrate that the proposed method outperforms the cutting-edge medical image segmentation methods on four publicly available datasets. We achieve the Dice Coefficient (DSC) of 0.9217, 0.9420, and 0.9224, 0.8824 on Kvasir-SEG, CVC-ClinicDB, 2018 Data Science Bowl dataset, and ISIC-2018 skin lesion segmentation challenge dataset respectively. We further conducted generalizability tests and achieved DSC of 0.7921 and 0.7575 on CVC-ClinicDB and Kvasir-SEG, respectively.
52 citations
TL;DR: Wang et al. as mentioned in this paper proposed a ciphertext-policy weighted attribute-based encryption (CP-WABE) scheme for the Internet of Health Things (IoHT), which supports both weighted attributes and any form of comparison of weighted attributes.
Abstract: The Internet of Health Things (IoHT) is a medical concept that describes uniquely identifiable devices connected to the Internet that can communicate with each other. As one of the most important components of smart health monitoring and improvement systems, the IoHT presents numerous challenges, among which cybersecurity is a priority. As a well-received security solution to achieve fine-grained access control, ciphertext-policy weighted attribute-based encryption (CP-WABE) has the potential to ensure data security in the IoHT. However, many issues remain, such as inflexibility, poor computational capability, and insufficient storage efficiency in attributes comparison. To address these issues, we propose a novel access policy expression method using 0-1 coding technology. Based on this method, a flexible and efficient CP-WABE is constructed for the IoHT. Our scheme supports not only weighted attributes but also any form of comparison of weighted attributes. Furthermore, we use offline/online encryption and outsourced decryption technology to ensure that the scheme can run on an inefficient IoT terminal. Both theoretical and experimental analyses show that our scheme is more efficient and feasible than other schemes. Moreover, security analysis indicates that our scheme achieves security against a chosen-plaintext attack.
51 citations
TL;DR: Deep Transfer Ensemble (DTE) as mentioned in this paper leverages the complementary feature views and diversity introduced by many different locally optimum solutions reached by individual networks through the randomization of hyper-parameters.
Abstract: Alzheimer's disease (AD) is one of the deadliest neurodegenerative diseases ailing the elderly population all over the world. An ensemble of Deep learning (DL) models can learn highly complicated patterns from MRI scans for the detection of AD by utilizing diverse solutions. In this work, we propose a computationally efficient, DL-architecture agnostic, ensemble of deep neural networks, named 'Deep Transfer Ensemble (DTE)' trained using transfer learning for the classification of AD. DTE leverages the complementary feature views and diversity introduced by many different locally optimum solutions reached by individual networks through the randomization of hyper-parameters. DTE achieves an accuracy of 99.05% and 85.27% on two independent splits of the large dataset for cognitively normal (NC) vs AD classification task. For the task of mild cognitive impairment (MCI) vs AD classification, DTE achieves 98.71% and 83.11% respectively on the two independent splits. It also performs reasonable on a small dataset consisting of only 50 samples per class. It achieved a maximum accuracy of 85% for NC vs AD on the small dataset. It also outperformed snapshot ensembles along with several other existing deep models from similar kind of previous works by other researchers.
37 citations
TL;DR: Wang et al. as mentioned in this paper proposed a scheme, that integrates blockchain-based confidentiality-privacy (CP) preserving scheme, CP-BDHCA, that operates in two phases, in the first phase, elliptic curve cryptographic (ECC)-based digital signature framework, HCA-ECC is proposed to establish a session key for secure communication among different healthcare entities.
Abstract: Healthcare big data (HBD) allows medical stakeholders to analyze, access, retrieve personal and electronic health records (EHR) of patients. Mostly, the records are stored on healthcare cloud and application (HCA) servers, and thus, are subjected to end-user latency, extensive computations, single-point failures, and security and privacy risks. A joint solution is required to address the issues of responsive analytics, coupled with high data ingestion in HBD and secure EHR access. Motivated from the research gaps, the paper proposes a scheme, that integrates blockchain (BC)-based confidentiality-privacy (CP) preserving scheme, CP-BDHCA, that operates in two phases. In the first phase, elliptic curve cryptographic (ECC)-based digital signature framework, HCA-ECC is proposed to establish a session key for secure communication among different healthcare entities. Then, in the second phase, a two-step authentication framework is proposed that integrates Rivest-Shamir-Adleman (RSA) and advanced encryption standard (AES), named as HCA-RSAE that safeguards the ecosystem against possible attack vectors. CP-BDAHCA is compared against existing HCA cloud applications in terms of parameters like response time, average delay, transaction and signing costs, signing and verifying of mined blocks, and resistance to DoS and DDoS attacks. We consider 10 BC nodes and create a real-world customized dataset to be used with SEER dataset. The dataset has 30,000 patient profiles, with 1000 clinical accounts. Based on the combined dataset the proposed scheme outperforms traditional schemes like AI4SAFE, TEE, Secret, and IIoTEED, with a lower response time. For example, the scheme has a very less response time of 300 ms in DDoS. The average signing cost of mined BC transactions is 3,34 seconds, and for 205 transactions, has a signing delay of 1405 ms, with improved accuracy of ≈ 12% than conventional state-of-the-art approaches.
35 citations
TL;DR: Wang et al. as mentioned in this paper proposed a three-dimensional convolutional attention neural network (3DCANN) for EEG emotion recognition, which is composed of spatio-temporal feature extraction module and EEG channel attention weight learning module.
Abstract: Since electroencephalogram (EEG) signals can truly reflect human emotional state, emotion recognition based on EEG has turned into a critical branch in the field of artificial intelligence. Aiming at the disparity of EEG signals in various emotional states, we propose a new deep learning model named three-dimension convolution attention neural network (3DCANN) for EEG emotion recognition in this paper. The 3DCANN model is composed of spatio-temporal feature extraction module and EEG channel attention weight learning module, which can extract the dynamic relation well among multi-channel EEG signals and the internal spatial relation of multi-channel EEG signals during continuous period time. In this model, the spatio-temporal features are fused with the weights of dual attention learning, and the fused features are input into the softmax classifier for emotion classification. In addition, we utilize SJTU Emotion EEG Dataset (SEED) to appraise the feasibility and effectiveness of the proposed algorithm. Finally, experimental results display that the 3DCANN method has superior performance over the state-of-the-art models in EEG emotion recognition.
35 citations
TL;DR: In this article , the authors classified brain tumors into three classes, namely glioma, meningioma and pituitary, using transfer learning model, and extracted features of the brain MRI images using a pre-trained CNN, i.e. GoogLeNet.
Abstract: In the healthcare research community, Internet of Medical Things (IoMT) is transforming the healthcare system into the world of the future internet. In IoMT enabled Computer aided diagnosis (CAD) system, the Health-related information is stored via the internet, and supportive data is provided to the patients. The development of various smart devices is interconnected via the internet, which helps the patient to communicate with a medical expert using IoMT based remote healthcare system for various life threatening diseases, e.g., brain tumors. Often, the tumors are predecessors to cancers, and the survival rates are very low. So, early detection and classification of tumors can save a lot of lives. IoMT enabled CAD system plays a vital role in solving these problems. Deep learning, a new domain in Machine Learning, has attracted a lot of attention in the last few years. The concept of Convolutional Neural Networks (CNNs) has been widely used in this field. In this paper, we have classified brain tumors into three classes, namely glioma, meningioma and pituitary, using transfer learning model. The features of the brain MRI images are extracted using a pre-trained CNN, i.e. GoogLeNet. The features are then classified using classifiers such as softmax, Support Vector Machine (SVM), and K-Nearest Neighbor (K-NN). The proposed model is trained and tested on CE-MRI Figshare and Harvard medical repository datasets. The experimental results are superior to the other existing models. Performance measures such as accuracy, specificity, and F1 score are examined to evaluate the performances of the proposed model.
34 citations
TL;DR: In this article , a hybrid intelligence-driven medical image recognition framework was proposed for remote patient diagnosis in IoMT, which combines deep learning with the conventional machine learning (CML)-based methods.
Abstract: In ear of smart cities, intelligent medical image recognition technique has become a promising way to solve remote patient diagnosis in IoMT. Although deep learning-based recognition approaches have received great development during the past decade, explainability always acts as a main obstacle to promote recognition approaches to higher levels. Because it is always hard to clearly grasp internal principles of deep learning models. In contrast, the conventional machine learning (CML)-based methods are well explainable, as they give relatively certain meanings to parameters. Motivated by the above view, this paper combines deep learning with the CML, and proposes a hybrid intelligence-driven medical image recognition framework in IoMT. On the one hand, the convolution neural network is utilized to extract deep and abstract features for initial images. On the other hand, the CML-based techniques are employed to reduce dimensions for extracted features and construct a strong classifier that output recognition results. A real dataset about pathologic myopia is selected to establish simulative scenario, in order to assess the proposed recognition framework. Results reveal that the proposal that improves recognition accuracy about two to three percent.
31 citations
TL;DR: In this paper , a review of the recent literature on technologies and methodologies for quantitative human gait analysis in the context of neurodegenerative diseases is presented, which can be of great support in both clinical diagnosis and severity assessment of these pathologies.
Abstract: This paper reviews the recent literature on technologies and methodologies for quantitative human gait analysis in the context of neurodegenerative diseases. The use of technological instruments can be of great support in both clinical diagnosis and severity assessment of these pathologies. In this paper, sensors, features and processing methodologies have been reviewed in order to provide a highly consistent work that explores the issues related to gait analysis. First, the phases of the human gait cycle are briefly explained, along with some non-normal gait patterns (gait abnormalities) typical of some neurodegenerative diseases. Then the paper reports the most common processing techniques for both feature selection and extraction and for classification and clustering. Finally, a conclusive discussion on current open problems and future directions is outlined.
31 citations
TL;DR: In this paper , an aspect-base sentiment analysis was conducted for USA, U.K., Canada, Turkey, France, Germany, Spain and Italy showing the approach of twitter users to vaccination and vaccine types during the COVID-19 period.
Abstract: Due to the COVID-19 pandemic, vaccine development and community vaccination studies are carried out all over the world. At this stage, the opposition to the vaccine seen in the society or the lack of trust in the developed vaccine is an important factor hampering vaccination activities. In this study, aspect-base sentiment analysis was conducted for USA, U.K., Canada, Turkey, France, Germany, Spain and Italy showing the approach of twitter users to vaccination and vaccine types during the COVID-19 period. Within the scope of this study, two datasets in English and Turkish were prepared with 928,402 different vaccine-focused tweets collected by country. In the classification of tweets, 4 different aspects (policy, health, media and other) and 4 different BERT models (mBERT-base, BioBERT, ClinicalBERT and BERTurk) were used. 6 different COVID-19 vaccines with the highest frequency among the datasets were selected and sentiment analysis was made by using Twitter posts regarding these vaccines. To the best of our knowledge, this paper is the first attempt to understand people's views about vaccination and types of vaccines. With the experiments conducted, the results of the views of the people on vaccination and vaccine types were presented according to the countries. The success of the method proposed in this study in the F1 Score was between 84% and 88% in datasets divided by country, while the total accuracy value was 87%.
29 citations
TL;DR: JAS-GAN as discussed by the authors proposes an adaptive attention cascade to automatically correlate the segmentation tasks of the unbalanced atrial targets, where the estimated LA acts as the attention map to adaptively focus on the small atrial scars roughly.
Abstract: Automated and accurate segmentations of left atrium (LA) and atrial scars from late gadolinium-enhanced cardiac magnetic resonance (LGE CMR) images are in high demand for quantifying atrial scars. The previous quantification of atrial scars relies on a two-phase segmentation for LA and atrial scars due to their large volume difference (unbalanced atrial targets). In this paper, we propose an inter-cascade generative adversarial network, namely JAS-GAN, to segment the unbalanced atrial targets from LGE CMR images automatically and accurately in an end-to-end way. Firstly, JAS-GAN investigates an adaptive attention cascade to automatically correlate the segmentation tasks of the unbalanced atrial targets. The adaptive attention cascade mainly models the inclusion relationship of the two unbalanced atrial targets, where the estimated LA acts as the attention map to adaptively focus on the small atrial scars roughly. Then, an adversarial regularization is applied to the segmentation tasks of the unbalanced atrial targets for making a consistent optimization. It mainly forces the estimated joint distribution of LA and atrial scars to match the real ones. We evaluated the performance of our JAS-GAN on a 3D LGE CMR dataset with 192 scans. Compared with the state-of-the-art methods, our proposed approach yielded better segmentation performance (Average Dice Similarity Coefficient (DSC) values of 0.946 and 0.821 for LA and atrial scars, respectively), which indicated the effectiveness of our proposed approach for segmenting unbalanced atrial targets.
TL;DR: Zhang et al. as mentioned in this paper proposed a deep learning framework to integrate a task-induced pyramid and attention generative adversarial network (TPA-GAN) with a pathwise transfer dense convolution network (PT-DCN) for imputation and classification of multimodal brain images.
Abstract: With the advance of medical imaging technologies, multimodal images such as magnetic resonance images (MRI) and positron emission tomography (PET) can capture subtle structural and functional changes of brain, facilitating the diagnosis of brain diseases such as Alzheimer's disease (AD). In practice, multimodal images may be incomplete since PET is often missing due to high financial costs or availability. Most of the existing methods simply excluded subjects with missing data, which unfortunately reduced the sample size. In addition, how to extract and combine multimodal features is still challenging. To address these problems, we propose a deep learning framework to integrate a task-induced pyramid and attention generative adversarial network (TPA-GAN) with a pathwise transfer dense convolution network (PT-DCN) for imputation and classification of multimodal brain images. First, we propose a TPA-GAN to integrate pyramid convolution and attention module as well as disease classification task into GAN for generating the missing PET data with their MRI. Then, with the imputed multimodal images, we build a dense convolution network with pathwise transfer blocks to gradually learn and combine multimodal features for final disease classification. Experiments are performed on ADNI-1/2 datasets to evaluate our method, achieving superior performance in image imputation and brain disease diagnosis compared to state-of-the-art methods.
TL;DR: In this article , the authors proposed a generic scheme VaCoChain that fuses blockchain and unmanned aerial vehicles (UAVs) underlying 5G communication services for timely vaccine distribution during novel coronavirus (COVID-19) and future pandemics.
Abstract: This paper proposes a generic scheme VaCoChain, that fuses blockchain (BC) and unmanned aerial vehicles (UAVs) underlying fifth-generation (5G) communication services for timely vaccine distribution during novel coronavirus (COVID-19) and future pandemics. The scheme offers 5G-tactile internet (5G-TI) based services for UAV communication networks (UAVCN) monitored through ground controller stations (GCs). 5G-TI enabled UAVCN supports real-time dense connectivity at ultra-low round-trip time (RTT) latency of [Formula: see text] and high availability of 99.99999%. Thus, it can support resilient vaccine distributions in a phased manner at government-designated nodal centers (NCs) with reduced round trip delays from vaccine production warehouses (VPW). Further, UAVCNs ensure minimizes human intervention and controls vaccine health conditions due to shorter trip times. Once vaccines are supplied at NCs warehouses, then the BC ensures timestamped documentation of vaccinated persons with chronology, auditability, and transparency of supply-chain checkpoints from VPW to NCs. Through smart contracts (SCs), priority groups can be formed for vaccination based on age, healthcare workers, and general commodities. In the simulation, for UAV efficacy, we have compared the scheme against fourth-generation (4G)-assisted long term evolution-advanced (LTE-A), orthogonal frequency division multiplexing (OFDM) channels, and traditional logistics for round-trip time (RTT) latency, logistics, and communication costs. In the BC setup, we have compared the scheme against the existing 5G-TI delivery scheme (Gupta et al.) for processing latency, packet losses, and transaction time. For example, in communication costs, the proposed scheme achieves an average improvement of 9.13 for block meta-information. For 4000 transactions, the proposed scheme has a communication latency of 16 s compared to 36 s. The packet loss is significantly reduced to 2.5% using 5G-TI compared to 16% in 4G-LTE-A. The proposed scheme has a computation cost of 1.6 ms and a communication cost of 157 bytes, which indicates the scheme efficacy against conventional approaches.
TL;DR: Based on the research of NSCLC targeted drugs, through the cross-fusion of synthetic biology and artificial intelligence, using the idea of bioengineering, the authors construct an artificial intelligence assisted medical system and propose a drug selection framework for the personalized selection of non-small cell lung cancer patients.
Abstract: According to statistics, in the 185 countries' 36 types of cancer, the morbidity and mortality of lung cancer take the first place, and non-small cell lung cancer (NSCLC) accounts for 85% of lung cancer (International Agency for Research on Cancer, 2018), (Bray et al. , 2018). Significantly in many developing countries, limited medical resources and excess population seriously affect the diagnosis and treatment of alung cancer patients. The 21st century is an era of life medicine, big data, and information technology. Synthetic biology is known as the driving force of natural product innovation and research in this era. Based on the research of NSCLC targeted drugs, through the cross-fusion of synthetic biology and artificial intelligence, using the idea of bioengineering, we construct an artificial intelligence assisted medical system and propose a drug selection framework for the personalized selection of NSCLC patients. Under the premise of ensuring the efficacy, considering the economic cost of targeted drugs as an auxiliary decision-making factor, the system predicts the drug effectiveness-cost then. The experiment shows that our method can rely on the provided clinical data to screen drug treatment programs suitable for the patient's conditions and assist doctors in making an efficient diagnosis.
TL;DR: In this article , the impact of regression algorithms on prediction accuracy in the brain age estimation frameworks have not been comprehensively evaluated; however, they have been shown to lead to more accurate brain age predictions in clinical settings.
Abstract: Machine learning (ML) algorithms play a vital role in the brain age estimation frameworks. The impact of regression algorithms on prediction accuracy in the brain age estimation frameworks have not been comprehensively evaluated. Here, we sought to assess the efficiency of different regression algorithms on brain age estimation. To this end, we built a brain age estimation framework based on a large set of cognitively healthy (CH) individuals ( N = 788) as a training set followed by different regression algorithms (22 different algorithms in total). We then quantified each regression-algorithm on independent test sets composed of 88 CH individuals, 70 mild cognitive impairment patients as well as 30 Alzheimer's disease patients. The prediction accuracy in the independent test set (i.e., CH set) varied in regression algorithms mean absolute error (MAE) from 4.63 to 7.14 yrs, R2 from 0.76 to 0.88. The highest and lowest prediction accuracies were achieved by Quadratic Support Vector Regression algorithm (MAE = 4.63 yrs, R2 = 0.88, 95% CI = [-1.26, 1.42]) and Binary Decision Tree algorithm (MAE = 7.14 yrs, R2 = 0.76, 95% CI = [-1.50, 2.62]), respectively. Our experimental results demonstrate that the prediction accuracy in brain age frameworks is affected by regression algorithms, indicating that advanced machine learning algorithms can lead to more accurate brain age predictions in clinical settings.
TL;DR: In this paper , the authors proposed an algorithm for simultaneous eye blink feature extraction and elimination from low-channel prefrontal EEG data, and compared the synergy of eye blink and EEG features before and after filtering by the proposed algorithm.
Abstract: Blink-related features derived from electroencephalography (EEG) have recently arisen as a meaningful measure of driver's cognitive state. Combined with band power features of low-channel prefrontal EEG data, blink-derived features enhance the detection of driver drowsiness. Yet, it remains unanswered whether synergy of combined blink and EEG band power features for the detection of driver drowsiness may be further boosted if a proper eye blink removal is also applied before EEG analysis. This paper proposes an algorithm for simultaneous eye blink feature extraction and elimination from low-channel prefrontal EEG data.Firstly, eye blink intervals (EBIs) are identified from the Fp1 EEG channel using variational mode extraction, and then blink-related features are derived. Secondly, the identified EBIs are projected to the rest of EEG channels and then filtered by a combination of principal component analysis and discrete wavelet transform. Thirdly, a support vector machine with 10-fold cross-validation is employed to classify alert and drowsy states from the derived blink and filtered EEG band power features.When compared the synergy of eye blink and EEG features before and after filtering by the proposed algorithm, a significant improvement in the mean accuracy of driver drowsiness detection was achieved (71.2% vs. 78.1%, p 0.05).This paper validates a novel view of eye blinks as both a source of information and artifacts in EEG-based driver drowsiness detection.
TL;DR: This paper proposes privacy infrastructure based on federated learning and blockchain technology that can effectively resolve the issue of large data silos and provide a shared model while preserving the privacy of data owners and information security and privacy analyses show that the proposed infrastructure is robust against information security related attacks.
Abstract: Internet of medical things (IoMT) has made it possible to collect applications and medical devices to improve healthcare information technology. Since the advent of the pandemic of coronavirus (COVID-19) in 2019, public health information has become more sensitive than ever. Moreover, different news items incorporated have resulted in differing public perceptions of COVID-19, especially on the social media platform and infrastructure. In addition, the unprecedented virality and changing nature of COVID-19 makes call centres to be likely overstressed, which is due to a lack of authentic and unregulated public media information. Furthermore, the lack of data privacy has restricted the sharing of COVID-19 information among health institutions. To resolve the above-mentioned limitations, this paper is proposing a privacy infrastructure based on federated learning and blockchain. The proposed infrastructure has the potentials to enhance the trust and authenticity of public media to disseminate COVID-19 information. Also, the proposed infrastructure can effectively provide a shared model while preserving the privacy of data owners. Furthermore, information security and privacy analyses show that the proposed infrastructure is robust against information security-related attacks.
TL;DR: The CNN-based skin disease classification merged with the federated learning approach is a breathtaking concept to classify human skin diseases while ensuring data security.
Abstract: Human skin disease, the most infectious dermatological ailment globally, is initially diagnosed by sight. Some clinical screening and dermoscopic analysis of skin biopsies and scrapings for accurate classification are medically compulsory. Classification of skin diseases using medical images is more challenging because of the complex formation and variant colors of the disease and data security concerns. Both the Convolution Neural Network (CNN) for classification and a federated learning approach for data privacy preservation show significant performance in the realm of medical imaging fields. In this paper, a custom image dataset was prepared with four classes of skin disease, a CNN model was suggested and compared with several benchmark CNN algorithms, and an experiment was carried out to ensure data privacy using a federated learning approach. An image augmentation strategy was followed to enlarge the dataset and make the model more general. The proposed model achieved a precision of 86%, 43%, and 60%, and a recall of 67%, 60%, and 60% for acne, eczema, and psoriasis. In the federated learning approach, after distributing the dataset among 1000, 1500, 2000, and 2500 clients, the model showed an average accuracy of 81.21%, 86.57%, 91.15%, and 94.15%. The CNN-based skin disease classification merged with the federated learning approach is a breathtaking concept to classify human skin diseases while ensuring data security.
TL;DR: In this article , a convolutional neural network (CNN) and recurrent neural networks (RNN) were used to perform segmentation and classification of 5 cardiac rhythms based on ECG recordings.
Abstract: This paper proposes a novel deep learning architecture involving combinations of Convolutional Neural Networks (CNN) layers and Recurrent neural networks (RNN) layers that can be used to perform segmentation and classification of 5 cardiac rhythms based on ECG recordings. The algorithm is developed in a sequence to sequence setting where the input is a sequence of five second ECG signal sliding windows and the output is a sequence of cardiac rhythm labels. The novel architecture processes as input both the spectrograms of the ECG signal as well as the heartbeats' signal waveform. Additionally, we are able to train the model in the presence of label noise. The model's performance and generalizability is verified on an external database different from the one we used to train. Experimental result shows this approach can achieve an average F1 scores of 0.89 (averaged across 5 classes). The proposed model also achieves comparable classification performance to existing state-of-the-art approach with considerably less number of training parameters.
TL;DR: Wang et al. as discussed by the authors proposed a hybrid neural network architecture consisting of convolutional, recurrent, and fully connected layers that operates directly on the raw photoplethysmogram (PPG) time series and provides BP estimation every 5 seconds.
Abstract: In this paper, we present a personalized deep learning approach to estimate blood pressure (BP) using the photoplethysmogram (PPG) signal. We propose a hybrid neural network architecture consisting of convolutional, recurrent, and fully connected layers that operates directly on the raw PPG time series and provides BP estimation every 5 seconds. To address the problem of limited personal PPG and BP data for individuals, we propose a transfer learning technique that personalizes specific layers of a network pre-trained with abundant data from other patients. We use the MIMIC III database which contains PPG and continuous BP data measured invasively via an arterial catheter to develop and analyze our approach. Our transfer learning technique, namely BP-CRNN-Transfer, achieves a mean absolute error (MAE) of 3.52 and 2.20 mmHg for SBP and DBP estimation, respectively, outperforming existing methods. Our approach satisfies both the BHS and AAMI blood pressure measurement standards for SBP and DBP. Moreover, our results demonstrate that as little as 50 data samples per person are required to train accurate personalized models. We carry out Bland-Altman and correlation analysis to compare our method to the invasive arterial catheter, which is the gold-standard BP measurement method.
TL;DR: Wang et al. as discussed by the authors proposed an explainable 3D Residual Attention Deep Neural Network (3D ResAttNet) for end-to-end learning from sMRI scans.
Abstract: Computer-aided early diagnosis of Alzheimer’s disease (AD) and its prodromal form mild cognitive impairment (MCI) based on structure Magnetic Resonance Imaging (sMRI) has provided a cost-effective and objective way for early prevention and treatment of disease progression, leading to improved patient care. In this work, we have proposed a novel computer-aided approach for early diagnosis of AD by introducing an explainable 3D Residual Attention Deep Neural Network (3D ResAttNet) for end-to-end learning from sMRI scans. Different from the existing approaches, the novelty of our approach is three-fold: 1) A Residual Self-Attention Deep Neural Network has been proposed to capture local, global and spatial information of MR images to improve diagnostic performance; 2) An explainable method using Gradient-based Localization Class Activation mapping (Grad-CAM) has been introduced to improve the interpretability of the proposed method; 3) This work has provided a full end-to-end learning solution for automated disease diagnosis. Our proposed 3D ResAttNet method has been evaluated on a large cohort of subjects from real datasets for two changeling classification tasks (i.e. Alzheimer’s disease (AD) vs. Normal cohort (NC) and progressive MCI (pMCI) vs. stable MCI (sMCI)). The experimental results show that the proposed approach has a competitive advantage over the state-of-the-art models in terms of accuracy performance and generalizability. The explainable mechanism in our approach is able to identify and highlight the contribution of the important brain parts (e.g., hippocampus, lateral ventricle and most parts of the cortex) for transparent decisions
TL;DR: A novel Transformer-based Generative Adversarial Network (RFormer) is proposed to restore the real degradation of clinical fundus images and significantly outperforms the state-of-the-art (SOTA) methods.
Abstract: Ophthalmologists have used fundus images to screen and diagnose eye diseases. However, different equipments and ophthalmologists pose large variations to the quality of fundus images. Low-quality (LQ) degraded fundus images easily lead to uncertainty in clinical screening and generally increase the risk of misdiagnosis. Thus, real fundus image restoration is worth studying. Unfortunately, real clinical benchmark has not been explored for this task so far. In this paper, we investigate the real clinical fundus image restoration problem. Firstly, We establish a clinical dataset, Real Fundus (RF), including 120 low- and high-quality (HQ) image pairs. Then we propose a novel Transformer-based Generative Adversarial Network (RFormer) to restore the real degradation of clinical fundus images. The key component in our network is the Window-based Self-Attention Block (WSAB) which captures non-local self-similarity and long-range dependencies. To produce more visually pleasant results, a Transformer-based discriminator is introduced. Extensive experiments on our clinical benchmark show that the proposed RFormer significantly outperforms the state-of-the-art (SOTA) methods. In addition, experiments of downstream tasks such as vessel segmentation and optic disc/cup detection demonstrate that our proposed RFormer benefits clinical fundus image analysis and applications.
TL;DR: The goal of FL-BETS is to identify and ensure the privacy preservation and fraud of data at various levels, such as local fog nodes and remote clouds, with minimum energy consumption and delay, and to satisfy the deadlines of healthcare workloads.
Abstract: These days, the usage of machine-learning-enabled dynamic Internet of Medical Things (IoMT) systems with multiple technologies for digital healthcare applications has been growing progressively in practice. Machine learning plays a vital role in the IoMT system to balance the load between delay and energy. However, the traditional learning models fraud on the data in the distributed IoMT system for healthcare applications are still a critical research problem in practice. The study devises a federated learning-based blockchain-enabled task scheduling (FL-BETS) framework with different dynamic heuristics. The study considers the different healthcare applications that have both hard constraint (e.g., deadline) and resource energy consumption (e.g., soft constraint) during execution on the distributed fog and cloud nodes. The goal of FL-BETS is to identify and ensure the privacy preservation and fraud of data at various levels, such as local fog nodes and remote clouds, with minimum energy consumption and delay, and to satisfy the deadlines of healthcare workloads. The study introduces the mathematical model. In the performance evaluation, FL-BETS outperforms all existing machine learning and blockchain mechanisms in fraud analysis, data validation, energy and delay constraints for healthcare applications.
TL;DR: In this article , a diabetic retinopathy generative adversarial network (DR-GAN) is proposed to synthesize high-resolution fundus images which can be manipulated with arbitrary grading and lesion information.
Abstract: Diabetic retinopathy (DR) is a complication of diabetes that severely affects eyes. It can be graded into five levels of severity according to international protocol. However, optimizing a grading model to have strong generalizability requires a large amount of balanced training data, which is difficult to collect, particularly for the high severity levels. Typical data augmentation methods, including random flipping and rotation, cannot generate data with high diversity. In this paper, we propose a diabetic retinopathy generative adversarial network (DR-GAN) to synthesize high-resolution fundus images which can be manipulated with arbitrary grading and lesion information. Thus, large-scale generated data can be used for more meaningful augmentation to train a DR grading and lesion segmentation model. The proposed retina generator is conditioned on the structural and lesion masks, as well as adaptive grading vectors sampled from the latent grading space, which can be adopted to control the synthesized grading severity. Moreover, a multi-scale spatial and channel attention module is devised to improve the generation ability to synthesize small details. Multi-scale discriminators are designed to operate from large to small receptive fields, and joint adversarial losses are adopted to optimize the whole network in an end-to-end manner. With extensive experiments evaluated on the EyePACS dataset connected to Kaggle, as well as the FGADR dataset, we validate the effectiveness of our method, which can both synthesize highly realistic ($1280 \times 1280$
TL;DR: In this article , an empirical intelligent agent (EIA) based on a unique Swarm-Neural Network (Swarm-NN) method is proposed to identify attackers in the edge-centric IoMT framework.
Abstract: The seamless integration of medical sensors and the Internet of Things (IoT) in smart healthcare has leveraged an intelligent Internet of Medical Things (IoMT) framework to detect the criticality of the patients. However, due to the limited storage capacity and computation power of the local IoT devices, patient's health data needs to transfer to remote computing devices for analysis, which can easily result in privacy leakage due to lack of control over the patient's health data and the vulnerability of the network for various types of attacks. Motivated by this, in this paper, an Empirical Intelligent Agent (EIA) based on a unique Swarm-Neural Network (Swarm-NN) method is proposed to identify attackers in the edge-centric IoMT framework. The major outcome of the proposed strategy is to identify the attacks during data transmission through a network and analyze the health data efficiently at the edge of the network with higher accuracy. The proposed Swarm-NN strategy is evaluated with a real-time secured dataset, namely the ToN-IoT dataset that collected Telemetry, Operating systems, and Network data for IoT applications and compares the performance over the standard classification models using various performance metrics. The test results demonstrate that the proposed Swarm-NN strategy achieves 99.5% accuracy over the ToN-IoT dataset.
TL;DR: In this article , the masking region of interest based on the attention mechanism was performed to improve the signal generators' precision, and the sine activation function was also used to retain more details when converting two signal domains.
Abstract: A non-invasive fetal electrocardiogram (FECG) is used to monitor the electrical pulse of the fetal heart. Decomposing the FECG signal from the maternal ECG (MECG) is a blind source separation problem, which is hard due to the low amplitude of the FECG, the overlap of R waves, and the potential exposure to noise from different sources. Traditional decomposition techniques, such as adaptive filters, require tuning, alignment, or pre-configuration, such as modeling the noise or desired signal to map the MECG to the FECG. The high correlation between maternal and fetal ECG fragments decreases the performance of convolution layers. Therefore, the masking region of interest based on the attention mechanism was performed to improve the signal generators' precision. The sine activation function was also used to retain more details when converting two signal domains. Three available datasets from the Physionet, including the A&D FECG, NI-FECG, and NI-FECG challenge, and one synthetic dataset using FECGSYN toolbox, were used to evaluate the performance. The proposed method could map an abdominal MECG to a scalp FECG with an average of 98% R-Square [CI 95%: 97%, 99%] as the goodness of fit on the A&D FECG dataset. Moreover, it achieved 99.7% F1-score [CI 95%: 97.8-99.9], 99.6% F1-score [CI 95%: 98.2%, 99.9%] and 99.3% F1-score [CI 95%: 95.3%, 99.9%] for fetal QRS detection on the A&D FECG, NI-FECG and NI-FECG challenge datasets, respectively. Also, the distortion was in the "very good" and "good" ranges. These results are comparable to the state-of-the-art results; thus, the proposed algorithm has the potential to be used for high-performance signal-to-signal conversion.
TL;DR: In this paper , a federated learning framework was proposed for epileptic seizure detection using a deep learning-based approach, which operates across a cluster of machines and achieved a sensitivity of 81.25%, a specificity of 82.00% and a geometric mean of 81% on the EPILEPSIAE dataset.
Abstract: Epilepsy is one of the most prevalent paroxystic neurological disorders. It is characterized by the occurrence of spontaneous seizures. About 1 out of 3 patients have drug-resistant epilepsy, thus their seizures cannot be controlled by medication. Automatic detection of epileptic seizures can substantially improve the patient's quality of life. To achieve a high-quality model, we have to collect data from various patients in a central server. However, sending the patient's raw data to this central server puts patient privacy at risk and consumes a significant amount of energy. To address these challenges, in this work, we have designed and evaluated a standard federated learning framework in the context of epileptic seizure detection using a deep learning-based approach, which operates across a cluster of machines. We evaluated the accuracy and performance of our proposed approach on the NVIDIA Jetson Nano Developer Kit based on the EPILEPSIAE database, which is one of the largest public epilepsy datasets for seizure detection. Our proposed framework achieved a sensitivity of 81.25%, a specificity of 82.00%, and a geometric mean of 81.62%. It can be implemented on embedded platforms that complete the entire training process in 1.86 hours using 344.34 mAh energy on a single battery charge. We also studied a personalized variant of the federated learning, where each machine is responsible for training a deep neural network (DNN) to learn the discriminative electrocardiography (ECG) features of the epileptic seizures of the specific person monitored based on its local data. In this context, the DNN benefitted from a well-trained model without sharing the patient's raw data with a server or a central cloud repository. We observe in our results that personalized federated learning provides an increase in all the performance metric, with a sensitivity of 90.24%, a specificity of 91.58%, and a geometric mean of 90.90%.
TL;DR: In this article , a subject-independent seizure predictor using Geometric Deep Learning (GDL) was proposed to synthesize subject-specific graphs using deep learning and achieved state-of-the-art performance using a one-hour early seizure prediction window.
Abstract: Recently, researchers in the biomedical community have introduced deep learning-based epileptic seizure prediction models using electroencephalograms (EEGs) that can anticipate an epileptic seizure by differentiating between the pre-ictal and interictal stages of the subject's brain. Despite having the appearance of a typical anomaly detection task, this problem is complicated by subject-specific characteristics in EEG data. Therefore, studies that investigate seizure prediction widely employ subject-specific models. However, this approach is not suitable in situations where a target subject has limited (or no) data for training. Subject-independent models can address this issue by learning to predict seizures from multiple subjects, and therefore are of greater value in practice. In this study, we propose a subject-independent seizure predictor using Geometric Deep Learning (GDL). In the first stage of our GDL-based method we use graphs derived from physical connections in the EEG grid. We subsequently seek to synthesize subject-specific graphs using deep learning. The models proposed in both stages achieve state-of-the-art performance using a one-hour early seizure prediction window on two benchmark datasets (CHB-MIT-EEG: 95.38% with 23 subjects and Siena-EEG: 96.05% with 15 subjects). To the best of our knowledge, this is the first study that proposes synthesizing subject-specific graphs for seizure prediction. Furthermore, through model interpretation we outline how this method can potentially contribute towards Scalp EEG-based seizure localization.
TL;DR: In this paper , a novel deep learning model called AACA-MLA-D-UNet is proposed to fully utilize the low-level detailed information and the complementary information encoded in different layers to accurately distinguish the vessels from the background with low model complexity.
Abstract: Automatic vessel segmentation in the fundus images plays an important role in the screening, diagnosis, treatment, and evaluation of various cardiovascular and ophthalmologic diseases. However, due to the limited well-annotated data, varying size of vessels, and intricate vessel structures, retinal vessel segmentation has become a long-standing challenge. In this paper, a novel deep learning model called AACA-MLA-D-UNet is proposed to fully utilize the low-level detailed information and the complementary information encoded in different layers to accurately distinguish the vessels from the background with low model complexity. The architecture of the proposed model is based on U-Net, and the dropout dense block is proposed to preserve maximum vessel information between convolution layers and mitigate the over-fitting problem. The adaptive atrous channel attention module is embedded in the contracting path to sort the importance of each feature channel automatically. After that, the multi-level attention module is proposed to integrate the multi-level features extracted from the expanding path, and use them to refine the features at each individual layer via attention mechanism. The proposed method has been validated on the three publicly available databases, i.e. the DRIVE, STARE, and CHASE _ DB1. The experimental results demonstrate that the proposed method can achieve better or comparable performance on retinal vessel segmentation with lower model complexity. Furthermore, the proposed method can also deal with some challenging cases and has strong generalization ability.
TL;DR: A lightweight privacy protection protocol based on a shared secret and weight mask, which is more accurate and efficient than federated learning for secure multiparty computing frameworks based on homomorphic encryption, which can work well in unstable edge computing environments such as smart healthcare.
Abstract: Edge intelligent computing is widely used in the fields, such as the Internet of Medical Things (IoMT), which has advantages, including high data processing efficiency, strong real-time performance and low network delay. However, there are many problems including privacy disclosure, limited calculation force, as well as scheduling and coordination issues. Federated learning can greatly improves training efficiency. However, due to the sensitive nature of the healthcare data, the aforementioned approach of transferring the patient’s data to the servers may create serious security and privacy issues. Therefore, this article proposes a Privacy Protection Scheme for Federated Learning under Edge Computing (PPFLEC). First of all, we propose a lightweight privacy protection protocol based on a shared secret and weight mask, which is based on a random mask scheme of secret sharing. It is more accurate and efficient than,homomorphic encryption. It can not only protect gradient privacy without losing model accuracy, but also resist equipment dropping and collusion attacks between devices. Second, we design an algorithm based on a digital signature and hash function, which achieves the integrity and consistency of the message, as well as resisting replay attacks. Finally, we propose a periodic average training strategy, compared with differential privacy to prove that our scheme is 40$\%$ faster in efficiency than in deferential privacy. Meanwhile, compared with federated learning, we can achieve the same efficiency under the condition of ensuring safety. Therefore, our scheme can work well in unstable edge computing environments such as smart healthcare.