Nema Salem
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Nema Salem has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Lina Muthanna
Effat University
2 shared publications - Rana Hassan
Effat University
2 shared publications - Abeer Mushtaq
Department of Primary and Secondary Healthcare, Lahore 54000, Pakistan
1 shared publication Abida Hussain
1 shared publication- Afshan Shah
Department of Electrical and Computer Engineering, COMSATS University Islamabad (CUI), Islamabad, Pakistan
1 shared publication - Ali Raza
Department of Primary and Secondary Healthcare, Lahore 54000, Pakistan
1 shared publication
Latest Publications
- Open Access
- 20 March 2026
- Published by Elsevier in SSRN Electronic Journal
ECG-Based Arrhythmia Detection for Smart City Healthcare: A Comparative Study of Dimensionality Reduction and Machine Learning Models
This paper presents an ECG-based arrhythmia classification framework aligned with smart healthcare applications in Smart City environments. The proposed system integrates dimensionality reduction with ML to enable efficient and scalable cardiac analytics suitable for resource-constrained and latency-sensitive urban health infrastructures. Principal Component Analysis (PCA) and Multi-Dimensional Scaling (MDS) are employed to obtain compact feature representations of 19 and 15 features, respectively, while preserving clinically relevant ECG characteristics. The reduced feature sets are evaluated using Logistic Regression, k-Nearest Neighbors (KNN), Decision Tree, Support Vector Machine (SVM), and Random Forest (RF) classifiers on ECG recordings from the MIT-BIH Arrhythmia Database. Experimental results show that RF and SVM achieve the highest overall performance across multiple arrhythmia classes, with weighted F1-scores of 0.9387 and 0.9491 on the validation set, respectively, indicating strong generalization and robustness. Comparative analysis highlights the adaptability of ensemble learning and the discriminative capability of margin-based classifiers under reduced feature representations, while KNN, Decision Tree, and Logistic Regression exhibit performance degradation under overlapping feature distributions. The proposed framework offers practical insights into accuracy–complexity trade-offs for scalable ECG analytics and supports the design of real-time cardiac monitoring pipelines within Smart City healthcare ecosystems, enabling efficient diagnostic support for connected hospitals, ambulatory monitoring, and wearable health platforms.
- 20 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessECG-Based Arrhythmia Detection for Smart City Healthcare: A Comparative Study of Dimensionality Reduction and Machine Learning Models
This paper presents an ECG-based arrhythmia classification framework aligned with smart healthcare applications in Smart City environments. The proposed system integrates dimensionality reduction with ML to enable efficient and scalable cardiac analytics suitable for resource-constrained and latency-sensitive urban health infrastructures. Principal Component Analysis (PCA) and Multi-Dimensional Scaling (MDS) are employed to obtain compact feature representations of 19 and 15 features, respectively, while preserving clinically relevant ECG characteristics. The reduced feature sets are evaluated using Logistic Regression, k-Nearest Neighbors (KNN), Decision Tree, Support Vector Machine (SVM), and Random Forest (RF) classifiers on ECG recordings from the MIT-BIH Arrhythmia Database. Experimental results show that RF and SVM achieve the highest overall performance across multiple arrhythmia classes, with weighted F1-scores of 0.9387 and 0.9491 on the validation set, respectively, indicating strong generalization and robustness. Comparative analysis highlights the adaptability of ensemble learning and the discriminative capability of margin-based classifiers under reduced feature representations, while KNN, Decision Tree, and Logistic Regression exhibit performance degradation under overlapping feature distributions. The proposed framework offers practical insights into accuracy–complexity trade-offs for scalable ECG analytics and supports the design of real-time cardiac monitoring pipelines within Smart City healthcare ecosystems, enabling efficient diagnostic support for connected hospitals, ambulatory monitoring, and wearable health platforms.
- 20 November 2025
- Published by Springer Nature
- p. 525-539
Blackhat Morphological Operations and Telea’s Inpainting for Effective Hair Artifact Removal in Dermatoscopic Images
- 20 November 2025
- Published by Springer Nature
- p. 525-539
Blackhat Morphological Operations and Telea’s Inpainting for Effective Hair Artifact Removal in Dermatoscopic Images
- 20 November 2025
- Published by Springer Nature
- p. 487-509
Cardiac Analysis with Advanced Feature Extraction and Reduction
- 20 November 2025
- Published by Springer Nature
- p. 487-509
Cardiac Analysis with Advanced Feature Extraction and Reduction
- 20 November 2025
- Published by Springer Nature
- p. 511-523
Segmentation of Skin Lesions in Dermatoscopic Images Using the Total Variation Chan–Vese Model
- 20 November 2025
- Published by Springer Nature
- p. 511-523
Segmentation of Skin Lesions in Dermatoscopic Images Using the Total Variation Chan–Vese Model
- 1 September 2025
- Published by Springer Nature
- p. 85-97
Efficiency of FL, ANFIS, and Classical DC Motor Controllers
- 1 September 2025
- Published by Springer Nature
- p. 85-97
Efficiency of FL, ANFIS, and Classical DC Motor Controllers
- Open Access
- 4 February 2025
- Published by Elsevier in Results in Engineering
- Vol. 25
Artificially intelligent detection of retinal pigment sign using P3S-Net for retinitis pigmentosa analysis
- 4 February 2025
- Published by Elsevier in Results in Engineering
- Vol. 25
Open AccessArtificially intelligent detection of retinal pigment sign using P3S-Net for retinitis pigmentosa analysis
- 26 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 90-96
Modeling and Analysis of a Thermoelectric Power Generator
Thermoelectric Generator (TEG) was developed using the Seebeck phenomenon. It consists of many thermocouples connected thermally in parallel and electrically in series to increase energy efficiency. TEGs instantly convert thermal energy to electrical energy with no rotating parts and are less likely to fail due to no moving parts. With the rising cost of fossil fuels and their negative impact on the atmosphere, it is time to consider TEGs as renewable energy sources with applications ranging from mW to W power. This study derives a mathematical model of a TEG module and validates it with MATLAB/SIMULINK.
- 26 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 90-96
Modeling and Analysis of a Thermoelectric Power Generator
Thermoelectric Generator (TEG) was developed using the Seebeck phenomenon. It consists of many thermocouples connected thermally in parallel and electrically in series to increase energy efficiency. TEGs instantly convert thermal energy to electrical energy with no rotating parts and are less likely to fail due to no moving parts. With the rising cost of fossil fuels and their negative impact on the atmosphere, it is time to consider TEGs as renewable energy sources with applications ranging from mW to W power. This study derives a mathematical model of a TEG module and validates it with MATLAB/SIMULINK.
- 4 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 23-28
Enhancing Cruise Performance Through PID Controller Tuned with Particle Swarm Optimization Technique
The Proportional-Integral-Derivative (PID) controller is a widely used feedback control mechanism in various applications, including automobile cruise control systems. The performance of a PID controller is highly dependent on the values of its tuning parameters, which can be challenging to determine in practice. Particle Swarm Optimization (PSO) has emerged as a powerful optimization algorithm that can tune the PID controller parameters for optimal performance. The PSO is a metaheuristic optimization algorithm inspired by the social behavior of birds and fish. The PSO-PID is a variant of the PID controller that employs PSO to optimize its tuning parameters. PSO-PID offers several advantages over traditional PID tuning methods, including improved accuracy, stability, and robustness. This paper briefly overviews the PSO-PID algorithm and its application to automobile cruise control systems. The paper discusses the key steps involved in PSO-PID tuning, including initialization, evaluation, update, and termination. It provides an example of how PSO-PID can achieve optimal vehicle speed control. The paper highlights the advantages of PSO-PID over traditional PID tuning methods. PSO-PID is a promising technology for cruise control systems and has the potential to significantly improve the safety, comfort, and efficiency of modern automobiles.
- 4 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 23-28
Enhancing Cruise Performance Through PID Controller Tuned with Particle Swarm Optimization Technique
The Proportional-Integral-Derivative (PID) controller is a widely used feedback control mechanism in various applications, including automobile cruise control systems. The performance of a PID controller is highly dependent on the values of its tuning parameters, which can be challenging to determine in practice. Particle Swarm Optimization (PSO) has emerged as a powerful optimization algorithm that can tune the PID controller parameters for optimal performance. The PSO is a metaheuristic optimization algorithm inspired by the social behavior of birds and fish. The PSO-PID is a variant of the PID controller that employs PSO to optimize its tuning parameters. PSO-PID offers several advantages over traditional PID tuning methods, including improved accuracy, stability, and robustness. This paper briefly overviews the PSO-PID algorithm and its application to automobile cruise control systems. The paper discusses the key steps involved in PSO-PID tuning, including initialization, evaluation, update, and termination. It provides an example of how PSO-PID can achieve optimal vehicle speed control. The paper highlights the advantages of PSO-PID over traditional PID tuning methods. PSO-PID is a promising technology for cruise control systems and has the potential to significantly improve the safety, comfort, and efficiency of modern automobiles.
- 4 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 01-06
Performance of LQR and PID Controllers for RS-550VC Motor Speed Enhancement
Motor speed control is a critical aspect of many industrial and commercial applications. This paper investigates the performance of two widely used control techniques, Linear Quadratic Regulator LQR and Proportional-Integral-Derivative PID, for motor speed control. We perform a simulation study using MATLAB to compare the performance of these two controllers in terms of their ability to track the reference speed profile. The simulation results show that LQR and PID controllers can achieve accurate speed tracking, but their performance characteristics differ significantly. LQR is better in the steady state and overshoot but slower than PID. Our study provides insights into the trade-offs between these two controllers and can guide the selection of an appropriate control technique for a given motor speed control application.
- 4 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 01-06
Performance of LQR and PID Controllers for RS-550VC Motor Speed Enhancement
Motor speed control is a critical aspect of many industrial and commercial applications. This paper investigates the performance of two widely used control techniques, Linear Quadratic Regulator LQR and Proportional-Integral-Derivative PID, for motor speed control. We perform a simulation study using MATLAB to compare the performance of these two controllers in terms of their ability to track the reference speed profile. The simulation results show that LQR and PID controllers can achieve accurate speed tracking, but their performance characteristics differ significantly. LQR is better in the steady state and overshoot but slower than PID. Our study provides insights into the trade-offs between these two controllers and can guide the selection of an appropriate control technique for a given motor speed control application.
- Open Access
- 11 May 2023
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 18 (5) , e0285629
Causal speech enhancement using dynamical-weighted loss and attention encoder-decoder recurrent neural network
Speech enhancement (SE) reduces background noise signals in target speech and is applied at the front end in various real-world applications, including robust ASRs and real-time processing in mobile phone communications. SE systems are commonly integrated into mobile phones to increase quality and intelligibility. As a result, a low-latency system is required to operate in real-world applications. On the other hand, these systems need efficient optimization. This research focuses on the single-microphone SE operating in real-time systems with better optimization. We propose a causal data-driven model that uses attention encoder-decoder long short-term memory (LSTM) to estimate the time-frequency mask from a noisy speech in order to make a clean speech for real-time applications that need low-latency causal processing. The encoder-decoder LSTM and a causal attention mechanism are used in the proposed model. Furthermore, a dynamical-weighted (DW) loss function is proposed to improve model learning by varying the weight loss values. Experiments demonstrated that the proposed model consistently improves voice quality, intelligibility, and noise suppression. In the causal processing mode, the LSTM-based estimated suppression time-frequency mask outperforms the baseline model for unseen noise types. The proposed SE improved the STOI by 2.64% (baseline LSTM-IRM), 6.6% (LSTM-KF), 4.18% (DeepXi-KF), and 3.58% (DeepResGRU-KF). In addition, we examine word error rates (WERs) using Google’s Automatic Speech Recognition (ASR). The ASR results show that error rates decreased from 46.33% (noisy signals) to 13.11% (proposed) 15.73% (LSTM), and 14.97% (LSTM-KF).
- 11 May 2023
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 18 (5) , e0285629
Open AccessCausal speech enhancement using dynamical-weighted loss and attention encoder-decoder recurrent neural network
Speech enhancement (SE) reduces background noise signals in target speech and is applied at the front end in various real-world applications, including robust ASRs and real-time processing in mobile phone communications. SE systems are commonly integrated into mobile phones to increase quality and intelligibility. As a result, a low-latency system is required to operate in real-world applications. On the other hand, these systems need efficient optimization. This research focuses on the single-microphone SE operating in real-time systems with better optimization. We propose a causal data-driven model that uses attention encoder-decoder long short-term memory (LSTM) to estimate the time-frequency mask from a noisy speech in order to make a clean speech for real-time applications that need low-latency causal processing. The encoder-decoder LSTM and a causal attention mechanism are used in the proposed model. Furthermore, a dynamical-weighted (DW) loss function is proposed to improve model learning by varying the weight loss values. Experiments demonstrated that the proposed model consistently improves voice quality, intelligibility, and noise suppression. In the causal processing mode, the LSTM-based estimated suppression time-frequency mask outperforms the baseline model for unseen noise types. The proposed SE improved the STOI by 2.64% (baseline LSTM-IRM), 6.6% (LSTM-KF), 4.18% (DeepXi-KF), and 3.58% (DeepResGRU-KF). In addition, we examine word error rates (WERs) using Google’s Automatic Speech Recognition (ASR). The ASR results show that error rates decreased from 46.33% (noisy signals) to 13.11% (proposed) 15.73% (LSTM), and 14.97% (LSTM-KF).