Electrocardiogram Waveforms Diagnosis based on Wavelet Representation and SqueezeNet Model

Managing atrial fibrillation in the CRT patient: Controversy or consensus?

A multi-layer convolutional neural network (MCNN) with hyperparameter optimization (HyperMCNN) is proposed for classifying human electrocardiograms (ECGs). For performance tests of the HyperMCNN, ECG recordings for patients with cardiac arrhythmia (ARR), congestive heart failure (CHF), and normal sinus rhythm (NSR) were obtained from three PhysioNet databases: MIT-BIH Arrhythmia Database, BIDMC Congestive Heart Failure Database, and MIT-BIH Normal Sinus Rhythm Database, respectively. The MCNN hyperparameters in convolutional layers included number of filters, filter size, padding, and filter stride. The hyperparameters in max-pooling layers were pooling size and pooling stride. Gradient method was also a hyperparameter used to train the MCNN model. Uniform experimental design approach was used to optimize the hyperparameter combination for the MCNN. In performance tests, the resulting 16-layer CNN with an appropriate hyperparameter combination (16-layer HyperMCNN) was used to distinguish among ARR, CHF, and NSR. The experimental results showed that the average correct rate and standard deviation obtained by the 16-layer HyperMCNN were superior to those obtained by a 16-layer CNN with a hyperparameter combination given by Matlab examples. Furthermore, in terms of performance in distinguishing among ARR, CHF, and NSR, the 16-layer HyperMCNN was superior to the 25-layer AlexNet, which was the neural network that had the best image identification performance in the ImageNet Large Scale Visual Recognition Challenge in 2012.

This paper proposes an approach for R-peak detection employing various signal processing techniques and abnormality detection in ECG signals involving classification into heart disorder categories like different arrhythmias and premature ventricular contractions using 2D Scalograms for Continuous Wavelet Transform (CWT) of signals and deep learning. For training the neural network, we have utilized the ECG Heartbeat Categorization Dataset from Kaggle online community. We utilized the ECG samples extracted from the MIT-BIH Arrhythmia Dataset and achieved a validation accuracy of 98.32%. The strategies employed for denoising the noisy ECG signal containing random noise, White Gaussian noise, 50 Hz AC Hum and high-frequency noise achieved a signal-to-noise ratio of 12.69 after filtering.

Deep learning assisted continuous wavelet transform-based spectrogram for the detection of chlorophyll content in potato leaves

Atrial fibrillation (AF) and congestive heart failure (CHF) are the most prevalent types of cardiovascular disorders as the leading cause of death due to delayed diagnosis. Early diagnosis of these cardiac conditions is possible by manually analyzing electrocardiogram (ECG) signals. However, manual diagnosis is complex, owing to the various characteristics of ECG signals. An accurate classification system for AF and CHF has the potential to save patient lives. Therefore, this study proposed an ECG signal classification system for AF and CHF using a one-dimensional convolutional neural network (1-D CNN) to provide a robust classification system performance. This study used ECG signal recording of AF, CHF, and NSR, which can be accessed on the Physionet website. A total of 5600 ECG signal segments were obtained from 56 subjects, divided into train sets from 42 subjects (N = 4200 ECG segments), and test sets from 14 subjects (N = 1400). We applied for leave-one-out cross-validation in training to select the best model. The proposed 1-D CNN algorithm successfully classified raw data of ECG signals into normal sinus rhythm (NSR), AF, and CHF by providing the highest classification accuracy of 99.643%, f1-score, recall, and precision of 0.996, respectively, with an AUC score of 0.999. The results showed that the proposed method extracted the ECG signal information directly without needing several preprocessing steps and feature extraction methods that potentially reduce the information contained in the ECG signals. Furthermore, the proposed method outperformed previous studies in classifying AF, CHF, and NSR. Therefore, this approach can be considered as an adjunct for medical personnel to diagnose AF, CHF, and NSR.

As an electrophysiologist, I am referred large numbers of patients with, or at risk of, arrhythmias as part of the heart failure complex. In this review I will concentrate on the role of pacemakers and defibrillators in heart failure, because these are my areas of greatest experience. Left ventricular assist devices and implantable hemodynamic monitors will not be reviewed. The first controversy for discussion is how to assess patients for suitability to receive these devices. Patients with heart failure are frequently fitted into a New York Heart Association (NYHA) class. The NYHA class is, however, a blunt instrument for decision-making that is based on a physician's or nurse's impression of how patients express their symptoms. Two patients with similar levels of symptomatology but with different responses to their symptoms may therefore receive different NYHA classifications and be candidates for different management approaches. Guidelines such as those from the European Society of Cardiology (ESC) offer valuable guidance on the use of cardiac resynchronization therapy (CRT).12 There are, however, issues in the use of CRT which the guidelines do not consider. Controversies that relate particularly to CRT include the identification of which patients benefit most from this technique and the characterization of its role in atrial fibrillation, in patients with bradycardia, and in non-responders. CRT has been available for approximately 10 years. An early meta-analysis of randomized controlled trials of CRT, including CONTAK-CD, InSync ICD, MIRACLE, and MUSTIC, showed that CRT reduced death from progressive heart failure by 51%, reduced heart failure hospitalization by 29%, and showed a trend to reduced all-cause mortality2 (Figures 12–3). Some of these trials included a defibrillator component, which made characterization of the effects of CRT alone problematic. Subsequent to these trials, the CARE-HF study was performed. CARE-HF observed that CRT produced a 37% reduction in all-cause mortality or unplanned hospitalization compared to medical therapy over 4 years (P < 0.0001)5 (Figure 4). All-cause mortality alone was reduced 36% by CRT when compared to optimal medical therapy (P = 0.0019). CARE-HT was the first major trial to demonstrate a definite improvement in all-cause mortality from CRT pacing. The COMPANION study included patients at higher risk than those in CARE-HF and investigated three treatment limbs – optimal pharmacological therapy, biventricular pacing, and biventricular defibrillation.3 COMPANION was not powered to detect a difference in outcome between the two types of biventricular device, but did appear to show that biventricular defibrillation was superior to medical therapy based on the secondary endpoint of all-cause mortality at 12 months (Figure 5). On the basis of these trials, the ESC guidelines recommend that “CRT using biventricular pacemakers can be considered [my italics] in patients with reduced ejection fraction and ventricular asynchrony (QRS width ≥120 ms) who remain symptomatic despite optimal medical therapy (NYHA III/IV) to improve symptoms, hospitalizations, and mortality”.12 From my viewpoint, this statement may require to be updated in the revised ESC guidelines and CRT should be performed much more commonly than presently. National guidelines from the Scottish Intercollegiate Guidelines Network (SIGN), which have been more produced recently (2007) than those from the ESC (2005), make a similar statement – namely, that CRT “should be considered” in patients with characteristics similar to those described in the ESC guidelines. The English National Institute for Clinical Excellence (NICE) Health Technology Appraisal of the place of CRT in heart failure (2007) states that: “CRT with a pacing device (CRT-P) is recommended as a treatment option for patients with heart failure who fulfill the following criteria. They are currently experiencing or have recently experienced NYHA class III-IV symptoms. They are in sinus rhythm either with a QRS duration 150 ms or longer … or with a QRS duration of 120–149 ms and mechanical dyssynchrony. They have a LVEF of 35% or less. They are receiving optimal pharmacological therapy.” The NICE guidelines therefore strongly recommend CRT for a precisely defined group of patients. The NICE guidelines also offer guidance on the use of CRT with a defibrillator (CRT-D). “CRT-D may be considered for people who fulfill the criteria for implantation of a CRT-P device and who separately fulfill the criteria for use of an ICD as recommended in NICE Technology Appraisal Guidance 95.” The latter criteria include wide QRS and poor LV function post-myocardial infarction. There is controversy over whether mechanical dyssynchrony is also required for the selection of patients for CRT. Considerable heterogeneity exists in the echocardiographic criteria adopted to define dyssynchrony, but a few small observational studies suggest that echocardiographic measurement of mechanical dyssynchrony may best identify patients who are likely to benefit from CRT.7 Further work is required in this area. What are the cost implications of using CRT-P and CRT-D? Based on data from the CARE-group using data from COMPANION, the incremental cost-effectiveness of CRT-P versus medical therapy was €7538 per quality-adjusted life-year (QALY), and the incremental cost-effectiveness of CRT-D versus CRT-P was €47 909.13 CRT-P was concluded to be relatively inexpensive, while CRT-D was of borderline cost-effectiveness. The patient's age comes into consideration in deciding who, among eligible patients, should receive CRT-D. In younger patients, CRT-D may be considered affordable in order to keep patients alive for as long as possible. In patients aged above 70 years, by contrast, other factors are more likely to cause death and the cost of CRT-D compared to CRT-P becomes prohibitive. The use of CRT in patients with atrial fibrillation (AF) is a contentious issue. The ESC guidelines do not specify a rhythm, but most other guidelines do specify the presence of sinus rhythm (SR) in their recommendations. Most trials have included only patients in SR, although the MUSTIC study included approximately equal numbers of patients in SR and AF. In MUSTIC, patients in SR improved significantly, whereas those in AF did not.4,8 For patients in SR, activation of the pacemaker was associated with an improvement in 6-min walking distance, which worsened when the device was inactivated. However, it is notable that only 37 of 64 patients in AF completed both limbs of the study. A long-term follow-up of patients in MUSTIC at 9 and 12 months found that all SR patients and 88% of AF patients were programmed to biventricular pacing.9 The two groups experienced similar magnitudes of improvement in walking distance, peak VO2, quality of life, NYHA class, and ejection fraction (Figure 6). Therefore, for every factor measured, AF patients appeared to do as well as SR patients in the long term. MUSTIC and some smaller studies represent the only trial evidence on which to base treatment in patients with AF. This raises the question of whether the potential benefits of CRT should be denied to AF patients in the absence of large-scale randomized studies. Opinions obtained from members of the Heart Rhythm UK suggest that most large centers are implanting CRT devices in a proportion of AF patients. There are two schools of thought on the type of AF patient who may benefit from CRT. In one school, CRT should be offered to the relatively small numbers of patients in whom rigorous rate control can be achieved by medical therapy and, if necessary, ablation of atrioventricular conduction to ensure biventricular pacing all of the time. According to the results of MUSTIC, this group may improve. This is called the “rate control” group. In the second school of thought, patients should be selected in whom stringent efforts will be made to restore and maintain SR. CRT may help by increasing the ejection fraction, lowering mitral regurgitation, and lowering left ventricular pressure. Cardioversion will be easier, and amiodarone and atrial catheter ablation may be added. This is called the “rhythm control” group. It is clear that the rate control group and the rhythm control groups encompass almost all patients with AF who fit CRT criteria. In other words, there is little agreement on which patients not to implant. Another controversial area is that, in every trial, there are patients classified as “non-responders”, who represent approximately 40% of the study population. In a proportion of these cases, however, the CRT device is stabilizing the patient who would otherwise deteriorate. Although the patient may complain of not feeling better, the device may actually be preventing deterioration or death. This situation illustrates the difficulty in measuring response. Other indices may need to be assessed, such brain natriuretic peptide levels, echocardiography/magnetic resonance imaging, and evidence of dyssynchrony, in order to identify patients who will benefit from CRT. CRT-D is an option in younger patients, especially those considered for future transplantation, while CRT-P may be provided in older patients to improve symptoms and reduce hospitalization. CRT-D may be considered in the presence of risk factors for sudden death, such as non-sustained ventricular tachycardia (VT) or T-wave alternans. The role of CRT-D in Class IV heart failure and AF requires clarification. Controversies that relate particularly to the use of implantable cardioverters-defibrillators (ICDs) include the identification of which patients will benefit from an ICD or from a CRT plus an ICD, whether the trial evidence for coronary artery disease can be extrapolated to dilated cardiomyopathy (DCM), and issues of deactivation at end of life. The ESC guidelines state that: “Implantation of an ICD in combination with biventricular pacing can be considered (my italics) in patients who remain symptomatic with severe heart failure (NYHA class III/IV) with LVEF [left ventricular ejection fraction] ≤35% and QRS duration >120 ms to improve morbidity or mortality”. “ICD therapy is recommended (my italics) to improve survival in patients who have survived cardiac arrest or who have sustained VT, which is either poorly tolerated or associated with reduced systolic LV function.” “ICD implantation is reasonable (my italics) in selected symptomatic patients with LVEF <30–35%, not within 40 days of a myocardial infarction, on optimal background therapy … to reduce sudden death.” The largest trial that has provided evidence to substantiate these statements is SCD-HeFT.1 SCD-HeFT randomized 2500 patients with symptomatic heart failure (NYHA II/III) and LVEF <35% to no antiarrhythmic therapy, amiodarone, or ICD with a 5-year follow-up. One half of the patients had an ischaemic etiology and one half had idiopathic DCM. Patients were relatively young (mean 60 years) and a proportion were overweight (mean 85 kg). Background medication use with an ACE inhibitor or angiotensin receptor blocker and beta-blocker was good, with rates approaching 90% and 80%, respectively, and use of spironolactone (31% at last follow-up) was reasonable. Loop diuretics were prescribed in the majority. Annual mortality over 5 years was 7.2% in the placebo group. Amiodarone had no influence on morality (hazard ratio [HR] 1.06. P = 0.529 vs placebo), but ICD significantly reduced the mortality rate (HR 0.77, P = 0.007 vs placebo) (Figure 7). Analysis of subgroups showed that the HR of mortality for ICD versus placebo was 0.54 in patients with NYHA II and 1.16 in those with NYHA class III, indicating that patients with Class II heart failure benefited while those with Class III tended to fare worse. Patients with an ischaemic or non-ischaemic etiology had HRs of 0.79 and 0.73, respectively, suggesting equal benefit from ICD. Further subgroup analyses indicated that patients with a low EF (≤30%) had the most to gain (HR 0.73, vs 1.08 for those with an EF >30%), as did those on beta-blocker therapy (HR 0.68, vs 0.92 for those not on beta-blocker therapy), which emphasizes that ICD therapy is not a substitute for good medical therapy. There have been problems in interpreting the outcomes of studies such as SCD-HeFT and MADIT-II.10 The reliance on subgroup analyses in these trials has reduced the power of these studies. Inaccuracies in measurement utilizing LVEF cutoff values and NYHA class has further hampered interpretation, and the relationship between the severity of the condition and the magnitude of improvement requires clarification. No “evidence base” exists on which to guide practice at end of life. Discussing these issues with the patient and relatives is difficult, but needs to be done openly and with sensitivity if any crisis or deterioration develops. Booklets and websites are a valuable source of information that patients can choose to look at, and base decisions on, at their own pace. In the UK, information for patients on websites is available from the British Heart Foundation (www.bhf.org) and the Arrhythmia Alliance (www.arrhythmiaalliance.org.uk). Primary prevention ICDs may be considered in stable heart failure patients who are already receiving optimal medical therapy. ICD should not be considered as a substitute for medical therapy. Ischaemic and non-ischaemic patients benefit equally from ICD, and the worse the LV function, the greater is the benefit. ICD may be considered as a bridge to transplant or, alternatively, as a means to keep the patient away from transplant.

Atrial fibrillation (AF) is common and highly variable in its clinical presentation and evolution. Iit causes substantial morbidity and mortality, including impaired quality of life, heart failure, systemic emboli, and stroke. An accurate diagnosis is important and should be distinguished from atrial flutter or other arrhythmias which involves the atrium. Management of patients with AF involves 3 objectives: rate control, rhythm control, and prevention of thromboembolism. A rate control strategy alone, without attempts at restoration or maintenance of sinus rhythm (SR), is reasonable in some patients with AF, especially those who are asymptomatic. In some circumstances, when the cause of AF is reversible, such as when AF occurs after cardiac surgery, no long-term therapy may be necessary. The CHADS2 scoring system can be used to risk stratify patients with nonvalvular AF to determine the need for warfarin. The risk of thromboembolism or stroke does not differ between pharmacological and electrical CV. Ablation of the AV conduction system and permanent pacing is an option for patients with rapid ventricular rates despite maximum medical therapy. However, there is growing concern about the negative effects of long-term RV pacing. Biventricular pacing may overcome many of the adverse hemodynamic effects associated with RV pacing. Pharmacological therapy to maintain SR should be considered in patients who have troublesome symptoms. Drugs should be used to decrease the frequency and duration of episodes, and to improve symptoms. AF recurrence while taking an antiarrhythmic drug is not indicative of treatment failure and does not necessitate a change in antiarrhythmic therapy. Antiarrhythmic drug choice is based on side effect profiles and the presence or absence of structural heart disease, heart failure, and hypertension. Catheter ablation for AF is currently considered a second-line therapy in highly symptomatic patients in whom one or more antiarrhythmic agents have failed. (J Karadiol Indones. 2010;31:187-95) Keywords: Atrial fibrillation, rate control, rhythm control, thromboembolisme, stroke, anticoagulation

This paper deals with design of an automatic detector for classification of selected cardiac arrhythmias. The proposed algorithms employ the continuous wavelet transform (CWT) combined with an analysis of its contour envelopes. The CWT was used in a detector of R-waves, to distinguish between normal and abnormal beats, and for detection of atrial premature contractions (APCs) and premature ventricular contractions (PVCs). The algorithm was validated by extensive testing on the MIT/BIH database. Searching for a local maximum in wavelet contour envelopes efficiently detects R- peaks. The overall accuracy of its detection tested on 48 half-hour signals is 99.5%. Two types of classifications were tested: 1. classification based on the contour envelope and the detection of significant points with overall accuracy 94.6%, 96.1% for the sinus rhythm (SR), 30.4% APCs, 71.2% PVCs and 2. the localization of maximum of square modulus of CWT coefficients in the area of QRS complex for the determination of PVCs between SR, right bundle branch block (RBBB), APC and other narrow complex arrhythmias with the accuracy 96.8%.

Our research introduces a framework that integrates edge computing, quantum transfer learning, and federated learning to revolutionize pain level assessment through ECG signal analysis. The primary focus lies in developing a robust, privacy-preserving system that accurately classifies pain levels (low, medium, and high) by leveraging the intricate relationship between pain perception and autonomic nervous system responses captured in ECG signals. At the heart of our methodology lies a signal processing approach that transforms one-dimensional ECG signals into rich, two-dimensional Continuous Wavelet Transform (CWT) images. These transformations capture both temporal and frequency characteristics of pain-induced cardiac variations, providing a comprehensive representation of autonomic nervous system responses to different pain intensities. Our framework processes these CWT images through a sophisticated quantum-classical hybrid architecture, where edge devices perform initial preprocessing and feature extraction while maintaining data privacy. The cornerstone of our system is a Quantum Convolutional Hybrid Neural Network (QCHNN) that harnesses quantum entanglement properties to enhance feature detection and classification robustness. This quantum-enhanced approach is seamlessly integrated into a federated learning framework, allowing distributed training across multiple healthcare facilities while preserving patient privacy through secure aggregation protocols. The QCHNN demonstrated remarkable performance, achieving a classification accuracy of 94.8% in pain level assessment, significantly outperforming traditional machine learning approaches.

An intelligent computer-aided diagnosis method for paroxysmal atrial fibrillation patients with nondiagnostic ECG signals

Patients with heart failure (HF) in sinus rhythm (SR) experience an increased incidence of thromboembolic events including stroke. Among patients with HF, high-quality evidence supports the use of oral anticoagulation when atrial fibrillation is present, but the benefit of anticoagulation in SR in the absence of other known indications for anticoagulation is unclear. In four randomized controlled trials (RCTs), warfarin did not improve a composite of clinical outcomes compared with aspirin or placebo in patients with HF with reduced ejection fraction (HFrEF) and SR. A recent RCT assessed the efficacy of the direct oral anticoagulant rivaroxaban versus placebo in patients with HFrEF (including mildly reduced ejection fraction), SR, and coronary artery disease. While rivaroxaban had a neutral effect on the primary composite outcome of myocardial infarction, stroke, or all-cause mortality, exploratory analyses revealed a significant reduction in strokes. It is thus possible that a subgroup of patients with HFrEF who are at high risk of stroke may benefit from anticoagulation. The challenge is to adequately identify this subgroup and to balance the potential benefit of anticoagulation with the risk of major bleeding. There is also an unmet need for evidence around anticoagulation in HF with preserved ejection fraction and SR. This review explores the current evidence around anticoagulation in patients with HF and SR, identifies challenges regarding outcome definitions and patient selection, and offers suggestions for future research.

Tachycardia induced cardiomyopathy (TIC), secondary to various tachyarrhythmias, is a reversible condition which can lead to cardiac enlargement and heart failure. The impairment of both structure and function of heart can be reverted completely or partially if tachyarrhythmias are ceased without delay. This study aimed to explore the clinical characteristics, therapeutic regimen and outcome of TIC in children. Clinical data of 12 children with TIC, who came from Peking University First Hospital from Feb. 2003 to Jun. 2009, were retrospectively analyzed and followed up. The echocardiogram data on admission were compared with those from 12 homochronous cases with idiopathic dilated cardiomyopathy matched with 12 TIC cases in age and gender. Atrial tachycardia is the commonest arrhythmia in 12 TIC cases (75%). Four cases underwent catheterization for radiofrequency ablation and all succeeded. The cardiac rhythm of 6 out of 8 cases treated with drugs became sinus rhythm after 3 days to 2 weeks antiarrhythmic drugs treatment. The remaining 2 cases still retained atrial rhythm, but the ventricular heart rates declined to normal. The left ventricular end-diastolic dimensions of the 12 cases were decreased compared with those of pretherapy [(37.5 ± 5.3) mm vs. (43.0 ± 5.7) mm, P < 0.01], and the left ventricular ejection fractions were increased [(60.5% ± 5.6%) vs. (33.7% ± 10.3%), P < 0.01], after (3.4 ± 2.3) months. In our (4.3 ± 2.4) year-follow-up, all cases were fine, except in one case the tachyarrhythmia relapsed because of discontinuation of the drug treatment by her parents. The left ventricular end-diastolic dimensions in 12 TIC cases were smaller than those of the 12 age- and gender-matched idiopathic dilated cardiomyopathy [(43.0 ± 5.7) mm vs. (54.8 ± 7.5) mm, t = 7.9, P < 0.01], and the ejection fractions were higher [(33.7% ± 10.3%) vs. (21.8% ± 7.5%), t = 3.7, P < 0.01]. The diagnosis of TIC should be considered for the children with tachycardia, cardiac enlargement and cardiac insufficiency. The degree of cardiac enlargement and cardiac insufficiency might be of value for the differential diagnosis between TIC and idiopathic dilated cardiomyopathy. The rhythm control and ventricular rates control could all result in a favorite therapeutic efficacy.

Atrial fibrillation (AF) occurs in one of four patients undergoing cardiac resynchronization therapy (CRT).-Without special therapy, the prognosis of AF patients with CRT has been generally worse than those in sinus rhythm. The importance of a high percentage of biventricular pacing (BIV%) was confirmed in a large study where the mortality was inversely associated with BIV% both in the presence of normal sinus and atrial paced rhythm and with AF. The greatest reduction in mortality was observed with BIV% >98%. Patients with BIV% >99.6% experienced a 24% reduction in mortality (p < 0.001) while those with BIV% <94.8% had a 19% increase in mortality. The optimal BIV% cut-point was 98.7%. This cutoff would appear mandatory but it would be best to approach 100%. Careful evaluation of device interrogation data upon which the BiV% is based is essential because the memorized data can vastly overestimate the percentage of truly resynchronized beats since it does not account for fusion and pseudofusion between intrinsic (not paced) and paced beats. The recently published randomized CERTIFY trial provides unequivocal proof of the value of AV junctional (AVJ) ablation in CRT patients with AF. This trial confirmed the favorable results of AVJ ablation by many other studies and two important meta-analyses and therefore established the firm recommendation that the procedure should be performed in most, if not all, patients with permanent AF as well as those with frequent and prolonged episodes of paroxysmal AF. Patients after AVJ have improved mortality with a mortality similar to those in sinus rhythm. The AVJ ablation procedure carries the theoretical risk of device failure and death in pacemaker dependent patients. An inappropriate first ICD shock for AF seems to increase mortality. Increased long-term mortality after an inappropriate shock may be due to the underlying atrial arrhythmia substrate as opposed to the effect of the shock itself.

Heart diseases are a major illness worldwide. There is a need for an accurate and reliable diagnosis procedure, which should not put heavy burden on the already overwhelmed medical staff, always available, and easily accessible, for people with high risk of heart diseases. Machine learning has the ability to learn from large amounts of data, and it may offer accurate and reliable diagnosis of new data. In this paper, two convolutional neural network (CCN) architectures have been evaluated, i.e AlexNet and GoogleNet, in order to help diagnosing four heart conditions: Arrhythmia (ARR), Atrial Fibrillation (AF), Congestive Heart Failure (CHF), and Normal Sinus Rhythm (NSR). A dataset of Electrocardiogram (ECG) Media files for heart related problems is fed into a deep learning (CNN) module to learn features and link them to corresponding labels. Results have showed that this technique is promising and could provide reliable solution to quick and reliable diagnosis of heart conditions, with an accuracy of 97.6%.

Detection and classification of partial discharge (PD) signal in XLPE cable is very important to find out the root cause of insulation failure. The purpose of this work is to improve the accuracy and effectiveness in classifying various types of PD signals including corona, surface, and internal discharges in a noisy environment using machine learning (ML) and deep learning (DL) techniques. For ML models, statistical features such as skewness, kurtosis, standard deviation, and variance are extracted from the PD signals and used for training. In contrast, the DL model focuses on continuous wavelet transforms (CWT) based scalogram images which highlight the resolution of the signal’s energy and are used for feature extraction in the classification study. Instead of using traditional phase-resolved partial discharge (PRPD) plots, this work introduces a novel approach to converting PD signals to scalogram images using CWT. These scalogram images provide a visual representation of the frequency components of different PD types and how they change over time. The wavelet-based scalogram images are used as input features for the proposed convolutional neural network (CNN) and state-of-the-art (SOTA) DL models, along with shallow ML classifiers for model training, validation, and testing. Along with this k-fold cross-validation and hyperparameter tuning are also employed to enhance the overall performance of the model. The classification results demonstrate that the proposed CNN model achieves up to 97.33% recognition accuracy and less computational complexity as compared to various ML and DL models.