MyMedR

Displaying publications 21 - 40 of 209 in total

Abstract:

Sort:

Fulltext A Hybrid Deep CNN Model for Abnormal Arrhythmia Detection Based on Cardiac ECG Signal

Ullah A, Rehman SU, Tu S, Mehmood RM, Fawad, Ehatisham-Ul-Haq M

Sensors (Basel), 2021 Feb 01;21(3).
PMID: 33535397 DOI: 10.3390/s21030951

Electrocardiogram (ECG) signals play a vital role in diagnosing and monitoring patients suffering from various cardiovascular diseases (CVDs). This research aims to develop a robust algorithm that can accurately classify the electrocardiogram signal even in the presence of environmental noise. A one-dimensional convolutional neural network (CNN) with two convolutional layers, two down-sampling layers, and a fully connected layer is proposed in this work. The same 1D data was transformed into two-dimensional (2D) images to improve the model's classification accuracy. Then, we applied the 2D CNN model consisting of input and output layers, three 2D-convolutional layers, three down-sampling layers, and a fully connected layer. The classification accuracy of 97.38% and 99.02% is achieved with the proposed 1D and 2D model when tested on the publicly available Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia database. Both proposed 1D and 2D CNN models outperformed the corresponding state-of-the-art classification algorithms for the same data, which validates the proposed models' effectiveness.

Matched MeSH terms: Electrocardiography*
Fulltext Electrocardiogram-Based Emotion Recognition Systems and Their Applications in Healthcare-A Review

Hasnul MA, Aziz NAA, Alelyani S, Mohana M, Aziz AA

Sensors (Basel), 2021 Jul 23;21(15).
PMID: 34372252 DOI: 10.3390/s21155015

Affective computing is a field of study that integrates human affects and emotions with artificial intelligence into systems or devices. A system or device with affective computing is beneficial for the mental health and wellbeing of individuals that are stressed, anguished, or depressed. Emotion recognition systems are an important technology that enables affective computing. Currently, there are a lot of ways to build an emotion recognition system using various techniques and algorithms. This review paper focuses on emotion recognition research that adopted electrocardiograms (ECGs) as a unimodal approach as well as part of a multimodal approach for emotion recognition systems. Critical observations of data collection, pre-processing, feature extraction, feature selection and dimensionality reduction, classification, and validation are conducted. This paper also highlights the architectures with accuracy of above 90%. The available ECG-inclusive affective databases are also reviewed, and a popularity analysis is presented. Additionally, the benefit of emotion recognition systems towards healthcare systems is also reviewed here. Based on the literature reviewed, a thorough discussion on the subject matter and future works is suggested and concluded. The findings presented here are beneficial for prospective researchers to look into the summary of previous works conducted in the field of ECG-based emotion recognition systems, and for identifying gaps in the area, as well as in developing and designing future applications of emotion recognition systems, especially in improving healthcare.

Matched MeSH terms: Electrocardiography*
Dynamic ECG features for atrial fibrillation recognition

Abdul-Kadir NA, Mat Safri N, Othman MA

Comput Methods Programs Biomed, 2016 Nov;136:143-50.
PMID: 27686711 DOI: 10.1016/j.cmpb.2016.08.021

BACKGROUND: Atrial fibrillation (AF) can cause the formation of blood clots in the heart. The clots may move to the brain and cause a stroke. Therefore, this study analyzed the ECG features of AF and normal sinus rhythm signals for AF recognition which were extracted by using a second-order dynamic system (SODS) concept.
OBJECTIVE: To find the appropriate windowing length for feature extraction based on SODS and to determine a machine learning method that could provide higher accuracy in recognizing AF.
METHOD: ECG features were extracted based on a dynamic system (DS) that uses a second-order differential equation to describe the short-term behavior of ECG signals according to the natural frequency (ω), damping coefficient, (ξ), and forcing input (u). The extracted features were windowed into 2, 3, 4, 6, 8, and 10 second episodes to find the appropriate windowing size for AF signal processing. ANOVA and t-tests were used to determine the significant features. In addition, pattern recognition machine learning methods (an artificial neural network (ANN) and a support vector machine (SVM)) with k-fold cross validation (k-CV) were used to develop the ECG recognition system.
RESULTS: Significant differences (p

Matched MeSH terms: Electrocardiography/methods*
Normal electrocardiogram after myocardial infarction

Ng WH, Ahmad Z

Med J Malaysia, 1978 Dec;33(2):128-32.
PMID: 755162
Matched MeSH terms: Electrocardiography*
Computer-aided diagnosis of congestive heart failure using ECG signals - A review

Jahmunah V, Oh SL, Wei JKE, Ciaccio EJ, Chua K, San TR, et al.

Phys Med, 2019 Jun;62:95-104.
PMID: 31153403 DOI: 10.1016/j.ejmp.2019.05.004

The heart muscle pumps blood to vital organs, which is indispensable for human life. Congestive heart failure (CHF) is characterized by the inability of the heart to pump blood adequately throughout the body without an increase in intracardiac pressure. The symptoms include lung and peripheral congestion, leading to breathing difficulty and swollen limbs, dizziness from reduced delivery of blood to the brain, as well as arrhythmia. Coronary artery disease, myocardial infarction, and medical co-morbidities such as kidney disease, diabetes, and high blood pressure all take a toll on the heart and can impair myocardial function. CHF prevalence is growing worldwide. It afflicts millions of people globally, and is a leading cause of death. Hence, proper diagnosis, monitoring and management are imperative. The importance of an objective CHF diagnostic tool cannot be overemphasized. Standard diagnostic tests for CHF include chest X-ray, magnetic resonance imaging (MRI), nuclear imaging, echocardiography, and invasive angiography. However, these methods are costly, time-consuming, and they can be operator-dependent. Electrocardiography (ECG) is inexpensive and widely accessible, but ECG changes are typically not specific for CHF diagnosis. A properly designed computer-aided detection (CAD) system for CHF, based on the ECG, would potentially reduce subjectivity and provide quantitative assessment for informed decision-making. Herein, we review existing CAD for automatic CHF diagnosis, and highlight the development of an ECG-based CAD diagnostic system that employs deep learning algorithms to automatically detect CHF.

Matched MeSH terms: Electrocardiography*
Fulltext Low tube voltage prospectively ECG-triggered coronary CT angiography: a systematic review of image quality and radiation dose

Tan SK, Yeong CH, Raja Aman RRA, Ng KH, Abdul Aziz YF, Chee KH, et al.

Br J Radiol, 2018 Jul;91(1088):20170874.
PMID: 29493261 DOI: 10.1259/bjr.20170874

OBJECTIVE: This study aimed (1) to perform a systematic review on scanning parameters and contrast medium (CM) reduction methods used in prospectively electrocardiography (ECG-triggered low tube voltage coronary CT angiography (CCTA), (2) to compare the achievable dose reduction and image quality and (3) to propose appropriate scanning techniques and CM administration methods.
METHODS: A systematic search was performed in PubMed, the Cochrane library, CINAHL, Web of Science, ScienceDirect and Scopus, where 20 studies were selected for analysis of scanning parameters and CM reduction methods.
RESULTS: The mean effective dose (HE) ranged from 0.31 to 2.75 mSv at 80 kVp, 0.69 to 6.29 mSv at 100 kVp and 1.53 to 10.7 mSv at 120 kVp. Radiation dose reductions of 38 to 83% at 80 kVp and 3 to 80% at 100 kVp could be achieved with preserved image quality. Similar vessel contrast enhancement to 120 kVp could be obtained by applying iodine delivery rate (IDR) of 1.35 to 1.45 g s-1 with total iodine dose (TID) of between 10.9 and 16.2 g at 80 kVp and IDR of 1.08 to 1.70 g s-1 with TID of between 18.9 and 20.9 g at 100 kVp.
CONCLUSION: This systematic review found that radiation doses could be reduced to a rate of 38 to 83% at 80 kVp, and 3 to 80% at 100 kVp without compromising the image quality. Advances in knowledge: The suggested appropriate scanning parameters and CM reduction methods can be used to help users in achieving diagnostic image quality with reduced radiation dose.

Matched MeSH terms: Electrocardiography*
Fulltext Bizarre ECG in head injury mimicking ventricular tachycardia

Loke YK, Lai VM, Tan MH, Gunn A

Singapore Med J, 1997 Apr;38(4):166-8.
PMID: 9269397

Bizarre electrocardiographic (ECG) changes were found in an 18-year-old girl who had a subdural haematoma following head trauma. The initial diagnosis was of ventricular tachycardia (VT) and she was treated with intravenous anti-arrhythmic drugs and electrical cardioversion, but to no effect. It was later concluded that the ECG appearances were not of a ventricular arrhythmia but were the result of the intracranial pathology. ECG abnormalities related to head injuries have been reported on many occasions, and our case report illustrates how this can create difficulties for the attending clinicians.

Matched MeSH terms: Electrocardiography*
The use of BLRP model for disaggregating daily rainfall affected by monsoon in Peninsular Malaysia

Zulkifli Yusop, Harisaweni, Fadhilah Yusof

Sains Malaysiana, 2016;45:87-97.

Rainfall intensity is the main input variable in various hydrological analysis and modeling. Unfortunately, the quality of rainfall data is often poor and reliable data records are available at coarse intervals such as yearly, monthly and daily. Short interval rainfall records are scarce because of high cost and low reliability of the measurement and the monitoring systems. One way to solve this problem is by disaggregating the coarse intervals to generate the short one using the stochastic method. This paper describes the use of the Bartlett Lewis Rectangular Pulse (BLRP) model. The method was used to disaggregate 10 years of daily data for generating hourly data from 5 rainfall stations in Kelantan as representative area affected by monsoon period and 5 rainfall stations in Damansara affected by inter-monsoon period. The models were evaluated on their ability to reproduce standard and extreme rainfall model statistics derived from the historical record over disaggregation simulation results. The disaggregation of daily to hourly rainfall produced monthly and daily means and variances that closely match the historical records. However, for the disaggregation of daily to hourly rainfall, the standard deviation values are lower than the historical ones. Despite the marked differences in the standard deviation, both data series exhibit similar patterns and the model adequately preserve the trends of all the properties used in evaluating its performances.

Matched MeSH terms: Electrocardiography
Prolonged QT predicts prognosis in COVID-19

Akhtar Z, Gallagher MM, Yap YG, Leung LWM, Elbatran AI, Madden B, et al.

Pacing Clin Electrophysiol, 2021 05;44(5):875-882.
PMID: 33792080 DOI: 10.1111/pace.14232

BACKGROUND: Coronavirus disease-2019 (COVID-19) causes severe illness and multi-organ dysfunction. An abnormal electrocardiogram is associated with poor outcome, and QT prolongation during the illness has been linked to pharmacological effects. This study sought to investigate the effects of the COVID-19 illness on the corrected QT interval (QTc).
METHOD: For 293 consecutive patients admitted to our hospital via the emergency department for COVID-19 between 01/03/20 -18/05/20, demographic data, laboratory findings, admission electrocardiograph and clinical observations were compared in those who survived and those who died within 6 weeks. Hospital records were reviewed for prior electrocardiograms for comparison with those recorded on presentation with COVID-19.
RESULTS: Patients who died were older than survivors (82 vs 69.8 years, p 455 ms (males) and >465 ms (females) (p = 0.028, HR 1.49 [1.04-2.13]), as predictors of mortality. QTc prolongation beyond these dichotomy limits was associated with increased mortality risk (p = 0.0027, HR 1.78 [1.2-2.6]).
CONCLUSION: QTc prolongation occurs in COVID-19 illness and is associated with poor outcome.

Matched MeSH terms: Electrocardiography
A novel automated diagnostic system for classification of myocardial infarction ECG signals using an optimal biorthogonal filter bank

Sharma M, Tan RS, Acharya UR

Comput Biol Med, 2018 11 01;102:341-356.
PMID: 30049414 DOI: 10.1016/j.compbiomed.2018.07.005

Myocardial infarction (MI), also referred to as heart attack, occurs when there is an interruption of blood flow to parts of the heart, due to the acute rupture of atherosclerotic plaque, which leads to damage of heart muscle. The heart muscle damage produces changes in the recorded surface electrocardiogram (ECG). The identification of MI by visual inspection of the ECG requires expert interpretation, and is difficult as the ECG signal changes associated with MI can be short in duration and low in magnitude. Hence, errors in diagnosis can lead to delay the initiation of appropriate medical treatment. To lessen the burden on doctors, an automated ECG based system can be installed in hospitals to help identify MI changes on ECG. In the proposed study, we develop a single-channel single lead ECG based MI diagnostic system validated using noisy and clean datasets. The raw ECG signals are taken from the Physikalisch-Technische Bundesanstalt database. We design a novel two-band optimal biorthogonal filter bank (FB) for analysis of the ECG signals. We present a method to design a novel class of two-band optimal biorthogonal FB in which not only the product filter but the analysis lowpass filter is also a halfband filter. The filter design problem has been composed as a constrained convex optimization problem in which the objective function is a convex combination of multiple quadratic functions and the regularity and perfect reconstruction conditions are imposed in the form linear equalities. ECG signals are decomposed into six subbands (SBs) using the newly designed wavelet FB. Following to this, discriminating features namely, fuzzy entropy (FE), signal-fractal-dimensions (SFD), and renyi entropy (RE) are computed from all the six SBs. The features are fed to the k-nearest neighbor (KNN). The proposed system yields an accuracy of 99.62% for the noisy dataset and an accuracy of 99.74% for the clean dataset, using 10-fold cross validation (CV) technique. Our MI identification system is robust and highly accurate. It can thus be installed in clinics for detecting MI.

Matched MeSH terms: Electrocardiography
Fulltext Basics of electrocardiogram (ECG) and its application in diagnosis of heart ailments: an educational series

Chodankar, Nagesh N., May, Honey Ohn, D’Souza, Urban John Arnold

Borneo Journal of Medical Sciences, 2018;12(1):3-22.
MyJurnal

Electrocardiogram (ECG) is a record of electrical activity of the heart. PQRST waves represent
the electrical activities of atria and ventricles. A complete three-dimensional electrical activity is
possible to be recorded using a 12-lead ECG. The normal and different routinely-met clinical ECG
are elaborated and discussed. This routine, normal and abnormal ECG, like arrhythmias and heart
block records as well as their clinical notes shall be educational information for the medical students.

Matched MeSH terms: Electrocardiography
Arrhythmia detection using deep convolutional neural network with long duration ECG signals

Yıldırım Ö, Pławiak P, Tan RS, Acharya UR

Comput Biol Med, 2018 11 01;102:411-420.
PMID: 30245122 DOI: 10.1016/j.compbiomed.2018.09.009

This article presents a new deep learning approach for cardiac arrhythmia (17 classes) detection based on long-duration electrocardiography (ECG) signal analysis. Cardiovascular disease prevention is one of the most important tasks of any health care system as about 50 million people are at risk of heart disease in the world. Although automatic analysis of ECG signal is very popular, current methods are not satisfactory. The goal of our research was to design a new method based on deep learning to efficiently and quickly classify cardiac arrhythmias. Described research are based on 1000 ECG signal fragments from the MIT - BIH Arrhythmia database for one lead (MLII) from 45 persons. Approach based on the analysis of 10-s ECG signal fragments (not a single QRS complex) is applied (on average, 13 times less classifications/analysis). A complete end-to-end structure was designed instead of the hand-crafted feature extraction and selection used in traditional methods. Our main contribution is to design a new 1D-Convolutional Neural Network model (1D-CNN). The proposed method is 1) efficient, 2) fast (real-time classification) 3) non-complex and 4) simple to use (combined feature extraction and selection, and classification in one stage). Deep 1D-CNN achieved a recognition overall accuracy of 17 cardiac arrhythmia disorders (classes) at a level of 91.33% and classification time per single sample of 0.015 s. Compared to the current research, our results are one of the best results to date, and our solution can be implemented in mobile devices and cloud computing.

Matched MeSH terms: Electrocardiography
Highly-sensitive graphene-based flexible pressure sensor platform

Muhammad Aniq Shazni, Lee MW, Lee HW

Sains Malaysiana, 2017;46:1155-1161.

In this work, graphene has been utilized as the sensing material for the development of a highly-sensitive flexible pressure sensor platform. It has been demonstrated that a graphene-based pressure sensor platform that is able to measure pressure change of up to 3 psi with a sensitivity of 0.042 psi-1 and a non-linearity of less than 1% has been accomplished. The developed device, which resides on a flexible platform, will be applicable for integration in continuous wearables health-care monitoring system for the measurement of blood pressure.

Matched MeSH terms: Electrocardiography
Fulltext A Shallow U-Net Architecture for Reliably Predicting Blood Pressure (BP) from Photoplethysmogram (PPG) and Electrocardiogram (ECG) Signals

Mahmud S, Ibtehaz N, Khandakar A, Tahir AM, Rahman T, Islam KR, et al.

Sensors (Basel), 2022 Jan 25;22(3).
PMID: 35161664 DOI: 10.3390/s22030919

Cardiovascular diseases are the most common causes of death around the world. To detect and treat heart-related diseases, continuous blood pressure (BP) monitoring along with many other parameters are required. Several invasive and non-invasive methods have been developed for this purpose. Most existing methods used in hospitals for continuous monitoring of BP are invasive. On the contrary, cuff-based BP monitoring methods, which can predict systolic blood pressure (SBP) and diastolic blood pressure (DBP), cannot be used for continuous monitoring. Several studies attempted to predict BP from non-invasively collectible signals such as photoplethysmograms (PPG) and electrocardiograms (ECG), which can be used for continuous monitoring. In this study, we explored the applicability of autoencoders in predicting BP from PPG and ECG signals. The investigation was carried out on 12,000 instances of 942 patients of the MIMIC-II dataset, and it was found that a very shallow, one-dimensional autoencoder can extract the relevant features to predict the SBP and DBP with state-of-the-art performance on a very large dataset. An independent test set from a portion of the MIMIC-II dataset provided a mean absolute error (MAE) of 2.333 and 0.713 for SBP and DBP, respectively. On an external dataset of 40 subjects, the model trained on the MIMIC-II dataset provided an MAE of 2.728 and 1.166 for SBP and DBP, respectively. For both the cases, the results met British Hypertension Society (BHS) Grade A and surpassed the studies from the current literature.

Matched MeSH terms: Electrocardiography
Fulltext Automatic identification of hypertension and assessment of its secondary effects using artificial intelligence: A systematic review (2013-2023)

Gudigar A, Kadri NA, Raghavendra U, Samanth J, Maithri M, Inamdar MA, et al.

Comput Biol Med, 2024 Apr;172:108207.
PMID: 38489986 DOI: 10.1016/j.compbiomed.2024.108207

Artificial Intelligence (AI) techniques are increasingly used in computer-aided diagnostic tools in medicine. These techniques can also help to identify Hypertension (HTN) in its early stage, as it is a global health issue. Automated HTN detection uses socio-demographic, clinical data, and physiological signals. Additionally, signs of secondary HTN can also be identified using various imaging modalities. This systematic review examines related work on automated HTN detection. We identify datasets, techniques, and classifiers used to develop AI models from clinical data, physiological signals, and fused data (a combination of both). Image-based models for assessing secondary HTN are also reviewed. The majority of the studies have primarily utilized single-modality approaches, such as biological signals (e.g., electrocardiography, photoplethysmography), and medical imaging (e.g., magnetic resonance angiography, ultrasound). Surprisingly, only a small portion of the studies (22 out of 122) utilized a multi-modal fusion approach combining data from different sources. Even fewer investigated integrating clinical data, physiological signals, and medical imaging to understand the intricate relationships between these factors. Future research directions are discussed that could build better healthcare systems for early HTN detection through more integrated modeling of multi-modal data sources.

Matched MeSH terms: Electrocardiography
Current methods in electrocardiogram characterization

Martis RJ, Acharya UR, Adeli H

Comput Biol Med, 2014 May;48:133-49.
PMID: 24681634 DOI: 10.1016/j.compbiomed.2014.02.012

The Electrocardiogram (ECG) is the P-QRS-T wave depicting the cardiac activity of the heart. The subtle changes in the electric potential patterns of repolarization and depolarization are indicative of the disease afflicting the patient. These clinical time domain features of the ECG waveform can be used in cardiac health diagnosis. Due to the presence of noise and minute morphological parameter values, it is very difficult to identify the ECG classes accurately by the naked eye. Various computer aided cardiac diagnosis (CACD) systems, analysis methods, challenges addressed and the future of cardiovascular disease screening are reviewed in this paper. Methods developed for time domain, frequency transform domain, and time-frequency domain analysis, such as the wavelet transform, cannot by themselves represent the inherent distinguishing features accurately. Hence, nonlinear methods which can capture the small variations in the ECG signal and provide improved accuracy in the presence of noise are discussed in greater detail in this review. A CACD system exploiting these nonlinear features can help clinicians to diagnose cardiovascular disease more accurately.

Matched MeSH terms: Electrocardiography/methods*
Person identification in irregular cardiac conditions using electrocardiogram signals

Sidek KA, Khalil I

Annu Int Conf IEEE Eng Med Biol Soc, 2011;2011:3772-5.
PMID: 22255160 DOI: 10.1109/IEMBS.2011.6090644

This paper presents a person identification mechanism in irregular cardiac conditions using ECG signals. A total of 30 subjects were used in the study from three different public ECG databases containing various abnormal heart conditions from the Paroxysmal Atrial Fibrillation Predicition Challenge database (AFPDB), MIT-BIH Supraventricular Arrthymia database (SVDB) and T-Wave Alternans Challenge database (TWADB). Cross correlation (CC) was used as the biometric matching algorithm with defined threshold values to evaluate the performance. In order to measure the efficiency of this simple yet effective matching algorithm, two biometric performance metrics were used which are false acceptance rate (FAR) and false reject rate (FRR). Our experimentation results suggest that ECG based biometric identification with irregular cardiac condition gives a higher recognition rate of different ECG signals when tested for three different abnormal cardiac databases yielding false acceptance rate (FAR) of 2%, 3% and 2% and false reject rate (FRR) of 1%, 2% and 0% for AFPDB, SVDB and TWADB respectively. These results also indicate the existence of salient biometric characteristics in the ECG morphology within the QRS complex that tends to differentiate individuals.

Matched MeSH terms: Electrocardiography/methods*
Fulltext Short QTc in epilepsy patients without cardiac symptoms

Teh HS, Tan HJ, Loo CY, Raymond AA

Med J Malaysia, 2007 Jun;62(2):104-8.
PMID: 18705439

Epilepsy patients have a higher mortality rate than the general population. Sudden unexpected death in epilepsy (SUDEP) is a major cause of mortality for these patients. The possibility of cardiac involvement in the pathogenesis of SUDEP has been suggested by many previous studies. This study compared the QT interval in epilepsy patients and normal controls, and identified the factors that affected the QT interval. Standard 12-lead ECGs were recorded from 70 consecutive epilepsy patients from the neurology clinic of HUKM and 70 age, race and gender matched controls. The mean QT interval corrected for heart rate (QTc) was calculated and compared. The mean QTc among the epilepsy patients was 0.401 +/- 0.027s. It was significantly shorter than the QTc (0.420 +/- 0.027s) in the control group (p<0.0005). Thirty five epilepsy patients (50%) and 17 matched controls (24.3%) had a mean QTc shorter than 0.40s (p=0.001). Among the epilepsy patients, the mean QTc did not significantly differ between patients in the duration (F=0.836, p=0.438) of the epilepsy, frequency (F=0.273, p=0.845) and types of seizures (p=0.633). There was no significant difference in the mean QTc between the epilepsy patients on different number of antiepileptic agents (F=0.444, p=0.643). Patients with cryptogenic epilepsy had a mean QTc of 0.392 +/- 0.029s, which was significantly shorter than patients with symptomatic epilepsy (QTc = 0.410 +/- 0.027s, p = 0.015). The mean QTc of the same subjects showed no significant interobserver difference (p=0.661). This study, for the first time, demonstrates that epilepsy patients have a significantly shorter QTc than controls, particularly in the subgroup of patients with cryptogenic epilepsy.
Study site: Neurology clinic, Pusat Perubatan Universiti Kebangsaan Malaysia (PPUKM), Kuala Lumpur, Malaysia

Matched MeSH terms: Electrocardiography*
Fulltext Evaluation of conventional and computerised monitoring of cardiac function in the coronary care unit

Ng KH

Med J Malaysia, 1983 Dec;38(4):289-93.
PMID: 6599984

One of the important functions of the Coronary Care Unit (CCU) is the continuous and intensive monitoring of cardiac function. To date, many monitoring techniques have been developed and tested. In this paper, both the conventional and computerised monitoring techniques are reviewed and evaluated. It is shown that a computerised system has several defirute advantages over the conventional system, e.g. lower false alarm rate, accurate and fast data processing, retrospective studies. However one also ought to be aware of the limitations,

Matched MeSH terms: Electrocardiography*
Automated detection of diabetic subject using pre-trained 2D-CNN models with frequency spectrum images extracted from heart rate signals

Yildirim O, Talo M, Ay B, Baloglu UB, Aydin G, Acharya UR

Comput Biol Med, 2019 10;113:103387.
PMID: 31421276 DOI: 10.1016/j.compbiomed.2019.103387

In this study, a deep-transfer learning approach is proposed for the automated diagnosis of diabetes mellitus (DM), using heart rate (HR) signals obtained from electrocardiogram (ECG) data. Recent progress in deep learning has contributed significantly to improvement in the quality of healthcare. In order for deep learning models to perform well, large datasets are required for training. However, a difficulty in the biomedical field is the lack of clinical data with expert annotation. A recent, commonly implemented technique to train deep learning models using small datasets is to transfer the weighting, developed from a large dataset, to the current model. This deep learning transfer strategy is generally employed for two-dimensional signals. Herein, the weighting of models pre-trained using two-dimensional large image data was applied to one-dimensional HR signals. The one-dimensional HR signals were then converted into frequency spectrum images, which were utilized for application to well-known pre-trained models, specifically: AlexNet, VggNet, ResNet, and DenseNet. The DenseNet pre-trained model yielded the highest classification average accuracy of 97.62%, and sensitivity of 100%, to detect DM subjects via HR signal recordings. In the future, we intend to further test this developed model by utilizing additional data along with cloud-based storage to diagnose DM via heart signal analysis.

Matched MeSH terms: Electrocardiography*

Filters

Please provide feedback to Administrator (afdal@afpm.org.my)

External Links