ABSTRACT
Wearable computers can be used in different domains including healthcare. However, due to suffering from challenges such as faults their applications may be limited in real practice. So, in designing wearable devices, designer must take into account fault tolerance techniques. This study aims to investigate the challenging issues of fault tolerance in wearable computing. For this purpose, different aspects of fault tolerance in wearable computing namely hardware, software, energy, and communication are studied; and state of the art research regarding each category is analysed. In this analysis, the performed works using the fault tolerance techniques are included in the form of 25 components and referred to as "fault tolerance plan". Using this fault tolerance plan and the appropriate profile, the fault tolerance of any wearable system can be evaluated. In this article, fault tolerances of several of the most prominent works conducted in the field of wearable computing were evaluated. The obtained results, with the medical profile, showed that only one wearable system had a fault tolerance of 91%, with the other systems having a fault tolerance of 24% or less. Also, the results obtained from evaluating these works, with the military profile, showed that only one wearable system had a fault tolerance of 76%, with the other systems having a fault tolerance of 19% or less. These mean that few studies have been conducted on the fault tolerance of wearable computing.
Subject(s)
Wearable Electronic Devices , Humans , Software , Equipment Design , Surveys and QuestionnairesABSTRACT
One of the most common causes of death worldwide is heart disease, including arrhythmia. Today, sciences such as artificial intelligence and medical statistics are looking for methods and models for correct and automatic diagnosis of cardiac arrhythmia. In pursuit of increasing the accuracy of automated methods, many studies have been conducted. However, in none of the previous articles, the relationship and structure between the heart leads have not been included in the model. It seems that the structure of ECG data can help develop the accuracy of arrhythmia detection. Therefore, in this study, a new structure of Electrocardiogram (ECG) data was introduced, and the Graph Convolution Network (GCN), which has the possibility of learning the structure, was used to develop the accuracy of cardiac arrhythmia diagnosis. Considering the relationship between the heart leads and clusters based on different ECG poles, a new structure was introduced. In this structure, the Mutual Information(MI) index was used to evaluate the relationship between the leads, and weight was given based on the poles of the leads. Weighted Mutual Information (WMI) matrices (new structure) were formed by R software. Finally, the 15-layer GCN network was adjusted by this structure and the arrhythmia of people was detected and classified by it. To evaluate the performance of the proposed new network, sensitivity, precision, specificity, accuracy, and confusion matrix indices were used. Also, the accuracy of GCN networks was compared by three different structures, including WMI, MI, and Identity. Chapman's 12-lead ECG Dataset was used in this study. The results showed that the values of sensitivity, precision, specificity, and accuracy of the GCN-WMI network with 15 intermediate layers were equal to 98.74%, 99.08%, 99.97% & 99.82%, respectively. This new proposed network was more accurate than the Graph Convolution Network-Mutual Information (GCN-MI) with an accuracy equal to 99.71% and GCN-Id with an accuracy equal to 92.68%. Therefore, utilizing this network, the types of arrhythmia were recognized and classified. Also, the new network proposed by the Graph Convolution Network-Weighted Mutual Information (GCN-WMI) was more accurate than those conducted in other studies on the same data set (Chapman). Based on the obtained results, the structure proposed in this study increased the accuracy of cardiac arrhythmia diagnosis and classification on the Chapman data set. Achieving such accuracy for arrhythmia diagnosis is a great achievement in clinical sciences.
Subject(s)
Arrhythmias, Cardiac , Electrocardiography , Neural Networks, Computer , Humans , Arrhythmias, Cardiac/diagnosis , Arrhythmias, Cardiac/physiopathology , Electrocardiography/methods , Algorithms , Signal Processing, Computer-AssistedABSTRACT
Classified as biomedical signal processing, cerebral signal processing plays a key role in human-computer interaction (HCI) and medical diagnosis. The motor imagery (MI) problem is an important research area in this field. Accurate solutions to this problem will greatly affect real-world applications. Most of the proposed methods are based on raw signal processing techniques. Known as prior knowledge, the structural-functional information and interregional connections can improve signal processing accuracy. It is possible to correctly perceive the generated signals by considering the brain structure (i.e. anatomical units), the source of signals, and the structural-functional dependence of different brain regions (i.e. effective connection) that are the semantic generators of signals. This study employed electroencephalograph (EEG) signals based on the activity of brain regions (cortex) and effective connections between brain regions based on dynamic causal modeling to solve the MI problem. EEG signals, as well as effective connections between brain regions to improve the interpretability of MI action, were fed into the architecture of Graph Convolutional Neural Network (GCN). The proposed model allowed GCN to extract more discriminative features. The results indicated that the proposed method was successful in developing a model with a MI detection accuracy of 93.73%.
ABSTRACT
The deep learning models such as AlexNet, VGG, and ResNet achieved a good performance in classifying the breast cancer histopathological images in BreakHis dataset. However, these models are not practically appropriate due to their computational complexity and too many parameters; as a result, they are rarely utilized on devices with limited computational resources. This paper develops a lightweight learning model based on knowledge distillation to classify the histopathological images of breast cancer in BreakHis. This method employs two teacher models based on VGG and ResNext to train two student models, which are similar to the teacher models in development but have fewer deep layers. In the proposed method, the adaptive joint learning approach is adopted to transfer the knowledge in the final-layer output of a teacher model along with the feature maps of its middle layers as the dark knowledge to a student model. According to the experimental results, the student model designed by ResNeXt architecture obtained the recognition rate 97.09% for all histopathological images. In addition, this model has â¼69.40 million fewer parameters, â¼0.93 G less GPU memory use, and 268.17 times greater compression rate than its teacher model. While in the student model the recognition rate merely dropped down to 1.75%. The comparisons indicated that the student model had a rather acceptable outputs compared with state-of-the-art methods in classifying the images of breast cancer in BreakHis.
Subject(s)
Breast Neoplasms , Data Compression , Humans , Female , Breast , StudentsABSTRACT
Cardiovascular diseases, like arrhythmia, as the leading causes of death in the world, can be automatically diagnosed using an electrocardiogram (ECG). The ECG-based diagnostic has notably resulted in reducing human errors. The main aim of this study is to increase the accuracy of arrhythmia diagnosis and classify various types of arrhythmias in individuals (suffering from cardiovascular diseases) using a novel graph convolutional network (GCN) benefitting from mutual information (MI) indices extracted from the ECG leads. In this research, for the first time, the relationships of 12 ECG leads measured using MI as an adjacency matrix were illustrated by the developed GCN and included in the ECG-based diagnostic method. Cross-validation methods were applied to select both training and testing groups. The proposed methodology was validated in practice by applying it to the large ECG database, recently published by Chapman University. The GCN-MI structure with 15 layers was selected as the best model for the selected database, which illustrates a very high accuracy in classifying different types of rhythms. The classification indicators of sensitivity, precision, specificity, and accuracy for classifying heart rhythm type, using GCN-MI, were computed as 98.45%, 97.89%, 99.85%, and 99.71%, respectively. The results of the present study and its comparison with other studies showed that considering the MI index to measure the relationship between cardiac leads has led to the improvement of GCN performance for detecting and classifying the type of arrhythmias, in comparison to the existing methods. For example, the above classification indicators for the GCN with the identity adjacency matrix (or GCN-Id) were reported to be 68.24%, 72.83%, 95.24%, and 92.68%, respectively.
Subject(s)
Cardiovascular Diseases , Neural Networks, Computer , Algorithms , Arrhythmias, Cardiac/diagnosis , Databases, Factual , Electrocardiography/methods , HumansABSTRACT
Magnification-independent (MI) classification is considered a promising method for detecting the histopathological images of breast cancer. However, it has too many parameters for real implementation due to dependence on input images in different magnification factors. In addition, magnification-dependent (MD) classification usually performs poorly on unseen samples, although it has lower input image sizes and fewer parameters. This paper proposes a novel method based on knowledge distillation (KD) to overcome the limitation of dissociation between MI classification and MD classification of breast cancer in histopathological images. The proposed KD method includes a pre-trained MI teacher model that is responsible for training an unprepared MD student model developed through only one magnification factor. In the proposed method, the decomposed feature maps of a teacher's intermediate layers are transferred as dark knowledge to a student. According to the experimental results, the student model developed through 40X images yielded accuracy rates of 99.41%, 99.26%, 99.14%, and 99.09% in response to unseen samples of 40X, 100X, 200X, and 400X images, respectively. Moreover, comparison results indicated the competitive performance of the proposed student model as opposed to the state-of-the-art method based on deep learning on BreakHis.
Subject(s)
Breast Neoplasms , Breast Neoplasms/pathology , Female , Humans , Knowledge Bases , Neural Networks, ComputerABSTRACT
The existing data mining solutions to identify risk factors associated with diseases are burdened with quite a few shortcomings. They usually use crisp partitions for numerical features and also do not use patient-specific profiles. These shortcomings create limitations for solving real problems. Discretizing a numerical feature through crisp partitions can also generate substantial partitioning errors, particularly for features whose values are closer to crisp boundaries. Since the normal range of each numerical feature varies according to the age, gender, and medical conditions of the patients, then ignoring these differences can undermine the accuracy of the extracted itemsets and rules. This paper presents a profile-based fuzzy association rule mining (PB-FARM) approach for the assessment of risk factors highly correlated with diseases. The proposed approach has three phases. Phase I involves creating profiles for patients based on their age, gender, and medical conditions, to determine a normal range of each numerical feature. Then fuzzy partitioning is done for all features (namely, numerical and categorical), and consequently, a structure, called FirstScan, is created. In Phase II, the FirstScan structure is utilized to mine for large fuzzy k-itemsets. Ultimately, in Phase III, the given k-itemsets are employed to generate fuzzy rules for associations between risk factors and diseases. To evaluate the performance of the proposed method the Z-Alizadeh Sani coronary artery disease (CAD) dataset, containing 303 records and 54 features, was used. The results show a positive correlation between typical chest pain and old age with the incidence of CAD. The comparisons made in this study showed that, firstly, the proposed algorithm has a higher partitioning accuracy than other methods, and secondly, it has a reasonably short execution time.
Subject(s)
Coronary Artery Disease , Data Mining , Algorithms , Causality , Fuzzy Logic , HumansABSTRACT
New biometric identification techniques are continually being developed to meet various applications. Electroencephalography (EEG) signals may provide a reasonable option for this type of identification due its unique features that overcome the lacks of other common methods. Currently, however, the processing load for such signals requires considerable time and labor. New methods and algorithms have attempted to reduce EEG processing time, including a reduction of the number of electrodes and segmenting the EEG data into its typical frequency bands. This work complements other efforts by proposing a genetic algorithm to reduce the number of necessary electrodes for measurements by EEG devices. Using a public EEG dataset of 109 subjects who underwent relaxation with eye-open and eye-closed stimuli, we aimed to determine the minimum set of electrodes required for optimum identification accuracy in each EEG sub-band of both stimuli. The results were encouraging and it was possible to accurately identify a subject using about 10 out of 64 electrodes. Moreover, higher frequency bands required a fewer number of electrodes for identification compared with lower frequency bands.
Subject(s)
Algorithms , Biometric Identification/methods , Electrodes , Electroencephalography/methods , Databases, Factual , HumansABSTRACT
Transcription factors positively and/or negatively impact gene expression by recruiting coregulatory factors, which interact through protein-protein binding. Here we demonstrate that mouse pancreas size and islet ß-cell function are controlled by the ATP-dependent Swi/Snf chromatin remodeling coregulatory complex that physically associates with Pdx1, a diabetes-linked transcription factor essential to pancreatic morphogenesis and adult islet cell function and maintenance. Early embryonic deletion of just the Swi/Snf Brg1 ATPase subunit reduced multipotent pancreatic progenitor cell proliferation and resulted in pancreas hypoplasia. In contrast, removal of both Swi/Snf ATPase subunits, Brg1 and Brm, was necessary to compromise adult islet ß-cell activity, which included whole-animal glucose intolerance, hyperglycemia, and impaired insulin secretion. Notably, lineage-tracing analysis revealed Swi/Snf-deficient ß-cells lost the ability to produce the mRNAs for Ins and other key metabolic genes without effecting the expression of many essential islet-enriched transcription factors. Swi/Snf was necessary for Pdx1 to bind to the Ins gene enhancer, demonstrating the importance of this association in mediating chromatin accessibility. These results illustrate how fundamental the Pdx1:Swi/Snf coregulator complex is in the pancreas, and we discuss how disrupting their association could influence type 1 and type 2 diabetes susceptibility.
Subject(s)
Cell Proliferation/physiology , Chromatin Assembly and Disassembly/physiology , DNA Helicases/metabolism , Homeodomain Proteins/metabolism , Insulin-Secreting Cells/metabolism , Nuclear Proteins/metabolism , Pancreas/metabolism , Trans-Activators/metabolism , Transcription Factors/metabolism , Animals , DNA Helicases/genetics , Gene Expression Regulation , Glucose Intolerance/genetics , Glucose Intolerance/metabolism , Homeodomain Proteins/genetics , Insulin/blood , Insulin-Secreting Cells/cytology , Mice , Mice, Transgenic , Nuclear Proteins/genetics , Pancreas/cytology , Trans-Activators/genetics , Transcription Factors/geneticsABSTRACT
Compression algorithm is an essential part of Telemedicine systems, to store and transmit large amount of medical signals. Most of existing compression methods utilize fixed transforms such as discrete cosine transform (DCT) and wavelet and usually cannot efficiently extract signal redundancy especially for non-stationary signals such as electroencephalogram (EEG). In this paper, we first propose learning-based adaptive transform using combination of DCT and artificial neural network (ANN) reconstruction technique. This adaptive ANN-based transform is applied to the DCT coefficients of EEG data to reduce its dimensionality and also to estimate the original DCT coefficients of EEG in the reconstruction phase. To develop a new near lossless compression method, the difference between the original DCT coefficients and estimated ones are also quantized. The quantized error is coded using Arithmetic coding and sent along with the estimated DCT coefficients as compressed data. The proposed method was applied to various datasets and the results show higher compression rate compared to the state-of-the-art methods.
Subject(s)
Data Compression/methods , Electroencephalography/methods , Neural Networks, Computer , Algorithms , Humans , Signal Processing, Computer-Assisted , TelemedicineABSTRACT
The development of sensors with the microelectromechanical systems technology expedites the emergence of new tools for human-computer interaction, such as inertial pens. These pens, which are used as writing tools, do not depend on a specific embedded hardware, and thus, they are inexpensive. Most of the available inertial pen character recognition approaches use the low-level features of inertial signals. This paper introduces a Persian/Arabic handwriting character recognition system for inertial-sensor-equipped pens. First, the motion trajectory of the inertial pen is reconstructed to estimate the position signals by using the theory of inertial navigation systems. The position signals are then used to extract high-level geometrical features. A new metric learning technique is then adopted to enhance the accuracy of character classification. To this end, a characteristic function is calculated for each character using a genetic programming algorithm. These functions form a metric kernel classifying all the characters. The experimental results show that the performance of the proposed method is superior to that of one of the state-of-the-art works in terms of recognizing Persian/Arabic handwriting characters.
Subject(s)
Algorithms , Handwriting , Image Processing, Computer-Assisted/methods , Pattern Recognition, Automated/methods , HumansABSTRACT
Advancements in computers and electronic technologies have led to the emergence of a new generation of efficient small intelligent systems. The products of such technologies might include Smartphones and wearable devices, which have attracted the attention of medical applications. These products are used less in critical medical applications because of their resource constraint and failure sensitivity. This is due to the fact that without safety considerations, small-integrated hardware will endanger patients' lives. Therefore, proposing some principals is required to construct wearable systems in healthcare so that the existing concerns are dealt with. Accordingly, this paper proposes an architecture for constructing wearable systems in critical medical applications. The proposed architecture is a three-tier one, supporting data flow from body sensors to cloud. The tiers of this architecture include wearable computers, mobile computing, and mobile cloud computing. One of the features of this architecture is its high possible fault tolerance due to the nature of its components. Moreover, the required protocols are presented to coordinate the components of this architecture. Finally, the reliability of this architecture is assessed by simulating the architecture and its components, and other aspects of the proposed architecture are discussed.
Subject(s)
Algorithms , Remote Sensing Technology/instrumentation , Telemedicine/instrumentation , Cloud Computing , Computer Communication Networks/instrumentation , Equipment Design , Equipment Failure , Humans , Reproducibility of ResultsABSTRACT
Automatic measurement and quantification of blood vessels' features and detection of vessel landmarks are key steps in the computer-aided diagnosis and diseases monitoring. This work proposes a novel and robust method for detecting vessel landmarks, i.e. bifurcation and crossovers, and measurement of different features, i.e. vessel orientation and vessel diameter as well as bifurcation angle, from the detected vessel network using simple and efficient local vessel pattern operator. The proposed method is applied to the publicly available DRIVE, STARE and ARIA databases and compared with existing state-of-the-art approaches. It shows higher accuracy in detection of vessel landmark and estimation of vessel features.