18 1735-2827 Iran University of Science and Technology (IUST) 3597 Deep Learning Editorial Note to the Special Issue: “Applications of Deep Learning in Electrical and Electronic Engineering” Faezmehr Ardalan b Fatemi Shahab c Daliri Mohammad Reza d b Biomedical Engineering Department, School of Electrical Engineering, Iran University of Science & Technology (IUST). c Biomedical Engineering Department, School of Electrical Engineering, Iran University of Science & Technology (IUST). d Biomedical Engineering Department, School of Electrical Engineering, Iran University of Science & Technology (IUST). 1 11 2024 20 4 1 7 08 12 2024 31 12 2024
3299 Speech Processing KurdSM: Transformer-Based Model for Kurdish Abstractive Text Summarization with an Annotated Corpus yamini pedram e daneshfar fatemeh f Ghorbani Abuzar g e Department of Computer Engineering, University of Kurdistan f Department of Computer Engineering, University of Kurdistan g Department of Computer Engineering, University of Isfahan 1 11 2024 20 4 8 22 26 05 2024 06 12 2024 With the exponential growth of unstructured data on the Web and social networks, extracting relevant information from multiple sources; has become increasingly challenging, necessitating the need for automated summarization systems. However, developing machine learning-based summarization systems largely depends on datasets, which must be evaluated to determine their usefulness in retrieving data. In most cases, these datasets are summarized with humans’ involvement. Nevertheless, this approach is inadequate for some low-resource languages, making summarization a daunting task. To address this, this paper proposes a method for developing the first abstractive text summarization corpus with human evaluation and automated summarization model for the Sorani Kurdish language. The researchers compiled various documents from information available on the Web (rudaw), and the resulting corpus was released publicly. A customized and simplified version of the mT5-base transformer was then developed to evaluate the corpus. The model's performance was assessed using criteria such as Rouge-1, Rouge-2, Rouge-L, N-gram novelty, manual evaluation and the results are close to reference summaries in terms of all the criteria. This unique Sorani Kurdish corpus and automated summarization model have the potential to pave the way for future studies, facilitating the development of improved summarization systems in low-resource languages. 3324 Biomedical Signal Processing Transforming Cardiac Care: Machine Learning in Heart Condition Prediction Using Phonocardiograms D’Souza Sandra h Reddy S Niranjan i Tarun Saikonda Krishna j P Sohan k acharya k aneesha l h Dept. of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, India-576104 i Dept. of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, India-576104 j Dept. of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, India-576104 k Manipal Institute of Technology, Manipal Academy of Higher Education (MAHE), Manipal, Karnataka, India l Dept. of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, India-576104 1 11 2024 20 4 23 32 08 06 2024 16 11 2024 The incidence of heart-related illnesses is on the rise worldwide. Heart diseases are primarily caused by a multitude of parameters, including high blood pressure, diabetes, and excessive cholesterol, which are controlled by poor dietary and lifestyle choices. The growth in cardiovascular diseases (CVD) is mostly due to several other behaviors, such as smoking, drinking, and sleeplessness. In the research, machine learning-based prediction methods work on the audio recordings of heartbeats known as phonocardiograms (PCG) to develop an algorithm that differentiates a normal healthy heart from an abnormal heart based on the heart sounds. The data set consists of 831 normal and 260 abnormal data, and the duration of each sample is 5 seconds. Features extracted from the data are up-sampled and applied to the logistic regression and random forest classification models. The developed models record a classification accuracy of 71% for logistic regression and 94% for the random forest model. Further, artificial neural networks (ANN) and Deep learning networks have been trained to improve performance and demonstrated an accuracy of 94.5%. 3331 Industrial Electronics Faster R-CNN and 3D reconstruction for handling tasks implementing a Scara robot Herrera-Benavidez Julian m Pachón-Suescún Cesar n Jimenez-Moreno Robinson o m Universidad Militar Nueva Granada n Universidad Militar Nueva Granada o Universidad Militar Nueva Granada 1 11 2024 20 4 33 40 13 06 2024 31 10 2024 This paper presents the design and results of using a deep learning algorithm for robotic manipulation in object handling tasks in a virtual industrial environment. The simulation tool used is V-REP and the environment corresponds to a production line based on a conveyor belt and a SCARA type robot manipulator. The main contribution of this work focuses on the integration of a depth camera located on the robot and the computation of the gripping coordinates by identifying and locating three different types of objects of interest with random locations on the conveyor belt, through a Faster R-CNN. The results show that the system manages to perform the indicated activities, obtaining a classification accuracy of 97.4% and a mean average precision of 0.93, which allowed a correct detection and manipulation of the objects. 3348 Signal Processing GPS Spoofing Detection using CAF Images and Neural Networks Based on the Proposed Peak Mapping Dimensionality Reduction Algorithm and TCNN Model Jahantab M. J. p Tohidi S. Mosavi Mohammad Reza Ayatollahi Ahmad p School of Electrical Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran 16846-13114, Iran. School of Electrical Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran 16846-13114, Iran. School of Electrical Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran 16846-13114, Iran. School of Electrical Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran 16846-13114, Iran. 1 11 2024 20 4 41 54 04 07 2024 03 11 2024 Global Positioning System (GPS)-based positioning has become an indispensable part of our daily lives. A GPS receiver calculates its distance from a satellite by measuring the signal reception delay. Then, after determining its position relative to at least four satellites, the receiver obtains its precise location in three dimensions. There is a fundamental flaw in this positioning system, namely that satellite signals at ground level are very weak and susceptible to interference in the bandwidth; therefore, even a slight interference can disrupt the GPS receiver. In this paper, spoofing detection based on the Cross Ambiguity Function (CAF) is used. Furthermore, a dimension reduction algorithm is proposed to improve the speed and performance of the detection process. The reduced-dimensional images are trained by a Convolutional Neural Network (CNN). Additionally, a modified CNN model as Transformed-CNN (TCNN) is presented to enhance accuracy in this paper. The simulation results show a 98.67% improvement in network training speed compared to images with original dimensions, a 1.16% improvement in detection accuracy compared to the baseline model with reduced dimensions, and a 9.83% improvement compared to the original dimensions in detecting spoofing, demonstrating the effectiveness of the proposed algorithm and model. 3384 Biological Modeling & Simulation Diagnosis of Coronary Heart Disease via Robust Artificial Neural Network Classifier by Adaptive Synthetic Sampling Approach Moradi Elahe Department of Electrical and Computer Engineering Yadegar-e-Imam Khomeini (RAH) Shahre Rey Branch, Islamic Azad University, Tehran, Iran 1 11 2024 20 4 55 67 27 07 2024 01 12 2024 With the intricate interplay between clinical and pathological data in coronary heart disease (CHD) diagnosis, there is a growing interest among researchers and healthcare providers in developing more accurate and reliable predictive methods. In this paper, we propose a new method entitled the robust artificial neural network classifier (RANNC) technique for the prediction of CHD. The dataset CHD in this paper has imbalanced data, and in addition, it has some outlier values. The dataset consists of information related to 4240 samples with 16 attributes. Due to the presence of outliers, a robust method has been used to scale the dataset. On the other hand, due to the imbalance of CHD data, three data balancing methods, including Random Over Sampling (ROS), Synthetic Minority Over Sampling Technique (SMOTE), and Adaptive Synthetic Sampling (ADASYN) approaches, have been applied to the CHD data set. Also, six artificial intelligence algorithms, including LRC, DTC, RFC, KNNC, SVC, and ANN, have been evaluated on the considered dataset with criteria such as precision, accuracy, recall, F1-score, and MCC. The RANNC, leveraging ADASYN to address data imbalance and outliers, significantly improved CHD diagnostic accuracy and the reliability of healthcare predictive models. It outperformed other artificial intelligence methods, achieving precision, accuracy, recall, F1-score, and MCC scores of 95.57%, 96.90%, 99.70%, 97.59%, and 93.42%, respectively. 3402 Neural Networks Advanced Photovoltaic Emulator with ANN-Based Modeling Using a DC-DC Push-Pull Converter and LQR Control with Current Observer HADJAISSA Aboubakeur BENMILOUD Mohammed AMEUR khaled BOUCHENAK HALIMA DIMEH Maria LACoSERE Laboratory, University of Laghouat, Algeria LACoSERE Laboratory, University of Laghouat, Algeria LACoSERE Laboratory, University of Laghouat, Algeria University of Laghouat University of Laghouat 1 11 2024 20 4 68 78 07 08 2024 22 10 2024 As solar photovoltaic power generation becomes increasingly widespread, the need for photovoltaic emulators (PVEs) for testing and comparing control strategies, such as Maximum Power Point Tracking (MPPT), is growing. PVEs allow for consistent testing by accurately simulating the behavior of PV panels, free from external influences like irradiance and temperature variations. This study focuses on developing a PVE model using deep learning techniques, specifically a Multi-Layer Perceptron (MLP) Artificial Neural Network (ANN) with backpropagation as the learning algorithm. The ANN is integrated with a DC-DC push-pull converter controlled via a Linear Quadratic Regulator (LQR) strategy. The ANN emulates the nonlinear characteristics of PV panels, generating precise reference currents. Additionally, the use of a single voltage sensor paired with a current observer enhances control signal accuracy and reduces the PVE system's hardware requirements. Comparative analysis demonstrates that the proposed LQR-based controller significantly outperforms conventional PID controllers in both steady-state error and response time. 3407 Optical Communications Machine Learning-Driven Adaptive Modulation for VLC-Enabled Medical Body Sensor Networks Bayat Rizi Reza Forouzan Amir R. Miramirkhani Farshad Sabahi Mohamad F. Dept. of Electrical Engineering, University of Isfahan Dept. of Electrical Engineering, University of Isfahan Department of Electrical and Electronics Engineering, Isik University Dept. of Electrical Engineering, University of Isfahan 1 11 2024 20 4 79 90 10 08 2024 14 12 2024 Visible Light Communication, a key optical wireless technology, offers reliable, high-bandwidth, and secure communication, making it a promising soloution for a variety of applications. Despite its many advantages, optical wireless communication faces challenges in medical environments due to fluctuating signal strength caused by patient movement. Smart transmitter structures can improve system performance by adjusting system parameters to the fluctuating channel conditions. The purpose of this research is to examine how adaptive modulation performs in a medical body sensor network system that uses visible light communication. The analysis focuses on various medical situations and investigates machine learning algorithms. The study compares adaptive modulation based on supervised learning with that based on reinforcement learning. The findings indicate that both approaches greatly improve spectral efficiency, emphasizing the significance of implementing link adaptation in visible light communication-based medical body sensor networks. The use of the Q-learning algorithm in adaptive modulation enables real-time training and enables the system to adjust to the changing environment without any prior knowledge about the environment. A remarkable improvement is observed for photodetectors on the shoulder and wrist since they experience more DC gain. 3428 Speech Recognition Virtual assistant robot for physical training exercises supervision Jimenez-Moreno Robinson Espitia Cubillos Anny Astrid Rodríguez Carmona Esperanza Universidad Militar Nueva Granada Universidad Militar Nueva Granada Universidad Militar Nueva Granada 1 11 2024 20 4 91 101 27 08 2024 24 10 2024 This document presents the design of a virtual robotic system for the supervision of physical training exercises, to be carried out in a closed environment, which only requires a computer equipment with a web camera. To do this, deep learning algorithms such as convolutional networks and short- and long-term memory networks are used to recognize voice commands and the user's video actions. A predefined dialogue template is used to guide a user's training cycle based on the execution of the exercises: push-ups, abdominal, jump or squat. The contribution of the work focuses on the integration of deep learning techniques to design and personalize virtual robotic assistants for everyday task. The results show a high level of accuracy by the virtual robot both in understanding the audio and in predicting the exercise to be performed, with a final accuracy value of 97.75% and 100%, respectively. 3433 Image Processing A Deep Learning Model for YOLOv9-based Human Abnormal Activity Detection: Violence and Non-Violence Classification 𝐒𝐚𝐥𝐞𝐡𝐢𝐧 𝐒𝐢𝐫𝐚𝐣𝐮𝐬 Rahman Shakila 𝐍𝐮𝐫 𝐌𝐨𝐡𝐚𝐦𝐦𝐚𝐝 𝐀𝐬𝐢𝐟 𝐀𝐡𝐦𝐚𝐝 𝐇𝐚𝐫𝐮𝐧 𝐌𝐨𝐡𝐚𝐦𝐦𝐚𝐝 𝐁𝐢𝐧 UDDIN JIA American International University, Bangladesh American International University, Bangladesh American International University, Bangladesh American International University, Bangladesh American International University, Bangladesh Woosong University 1 11 2024 20 4 102 114 05 09 2024 16 11 2024 Abnormal activity detection is crucial for video surveillance and security systems, aiming to identify behaviors that deviate from normal patterns and may indicate threats or incidents such as theft, vandalism, accidents, and aggression. Timely recognition of these activities enhances public safety across various environments, including transportation hubs, public spaces, workplaces, and homes. In this study, we focus on detecting violent and non-violent activities of humans using a YOLOv9-based deep learning model considering the above issues. A diverse dataset has been built of 9,341 images from various platforms, and then the dataset has been pre-processed, i.e., augmentation, resizing, and annotating. After pre-processing, the proposed model has been trained which demonstrated strong performance, achieving an F1 score of 95% during training for 150 epochs. It was also trained for 200 epochs, but early stopping was applied at 148 epochs as there was no significant improvement in the results. Finally, the results of the YOLOv9-based model have been analyzed with other baseline models (YOLOv5, YOLOv7, YOLOv8, and YOLOv10) and it performed better compared with others. 3459 Machine Learning Deep Learning Integration in PAPR Reduction in 5G Filter Bank Multicarrier Systems Moharam Mohamed Hussien wafik AYA W. Assistant Professor at Misr University for Science and Technology, Electronics and Communications Engineering Department, Giza, Egypt Cyber Security Engineer graduate from Misr University for Science and Technology, Electronics and Communications Engineering Department, Giza, Egypt. 1 11 2024 20 4 115 125 22 09 2024 15 12 2024 High peak-to-average power ratio (PAPR) has been a major drawback of Filter bank Multicarrier (FBMC) in the 5G system. This research aims to calculate the PAPR reduction associated with the FBMC system. This research uses four techniques to reduce PAPR. They are classical tone reservation (TR). It combines tone reservation with sliding window (SW-TR). It also combines them with active constellation extension (TRACE) and with deep learning (TR-Net). TR-net decreases the greatest PAPR reduction by around 8.6 dB compared to the original value. This work significantly advances PAPR reduction in FBMC systems by proposing three hybrid methods, emphasizing the deep learning-based TRNet technique as a groundbreaking solution for efficient, distortion-free signal processing. 3463 Biomedical Signal Processing Tuberculosis Classification using SVM and Modified CNN BABU N SRINIVAS S SHASHIKIRAN JAYANTHI M N Rajani PALANISWAMY K M KUSHALATHA M R Department of Electronics and Communication Engineering, New Horizon College of Engineering, Bangalore, Karnataka India. Department of Electronics and Communication Engineering, New Horizon College of Engineering, Bangalore, Karnataka India. Department of Electronics and Communication Engineering, New Horizon College of Engineering, Bangalore, Karnataka India. Department of Electronics and Communication Engineering, Nitte Meenakshi Institute of Technology, Bengaluru, Karnataka, India. Department of Electronics and Communication Engineering, Dr T Thimmaiah Institute of Technology, KGF, Karnataka, India. Department of Electronics and Communication Engineering, Nitte Meenakshi Institute of Technology, Bengaluru, Karnataka, India. 1 11 2024 20 4 126 133 23 09 2024 01 12 2024 Tuberculosis (TB) is a dangerous disease caused by mycobacterium leads to mortality. Early detection and identification of tuberculosis is crucial for managing tuberculosis infections. Recent technological improvements use a machine learning-based SVM and Modified CNN to identify specific diseases more accurately, as demonstrated in this research. The modified CNN's improved feature extraction and classification accuracy are maintained throughout construction. To obtain good performance a TBX11K publicly accessible dataset is used it consists of 11000 images of which 4600 chest x-ray (CXR) images are considered in this research, and the suggested model is verified. This approach significantly increases the accuracy of categorizing TB symptoms.  The PCA in this system locates the elements and extracts a large amount of variance technique applied to the full chest radiograph for pulmonary tuberculosis identification accuracy using SVM is 93.14% and modified CNN 96.72% respectively. When it comes to helping radiologists diagnose patients and public health professionals screen for tuberculosis in places where the disease is endemic, the proposed system SVM and modified CNN perform better than existing methods. 3472 Machine Learning Grasp Area Detection for 3D Object using Enhanced Dynamic Graph Convolutional Neural Network Merrikhi Haniye Ebrahimnezhad Hossein Computer Vision Res. Lab., Electrical Engineering Faculty, Sahand University of Technology, Tabriz, Iran. Computer Vision Res. Lab., Electrical Engineering Faculty, Sahand University of Technology, Tabriz, Iran. 1 11 2024 20 4 134 146 28 09 2024 22 12 2024 Robots have become integral to modern society, taking over both complex and routine human tasks. Recent advancements in depth camera technology have propelled computer vision-based robotics into a prominent field of research. Many robotic tasks—such as picking up, carrying, and utilizing tools or objects—begin with an initial grasping step. Vision-based grasping requires the precise identification of grasp locations on objects, making the segmentation of objects into meaningful components a crucial stage in robotic grasping. In this paper, we present a system designed to detect the graspable parts of objects for a specific task. Recognizing that everyday household items are typically grasped at certain sections for carrying, we created a database of these objects and their corresponding graspable parts. Building on the success of the Dynamic Graph CNN (DGCNN) network in segmenting object components, we enhanced this network to detect the graspable areas of objects. The enhanced network was trained on the compiled database, and the visual results, along with the obtained Intersection over Union (IoU) metrics, demonstrate its success in detecting graspable regions. It achieved a grand mean IoU (gmIoU) of 92.57% across all classes, outperforming established networks such as PointNet++ in part segmentation for this dataset. Furthermore, statistical analysis using analysis of variance (ANOVA) and T-test validates the superiority of our method. 3486 Image Processing Performance Comparison of Facial Emotion Recognition: Introducing a Model within the Driver Assistance Framework based on Deep Learning with LBP Feature Extraction for In-Vehicle Applications Ghasemi Ehsan Razavi Seyyed Mohammad Mohamadzadeh Sajad Department of Electrical and Computer Engineering, University of Birjand Department of Electrical and Computer Engineering, University of Birjand Department of Electrical and Computer Engineering, University of Birjand 1 11 2024 20 4 147 161 07 10 2024 28 12 2024 This study proposes a descriptor-based approach combined with deep learning, which recognizes facial emotions for safe driving.  Paying attention to the driver's facial expressions is crucial to address the increasing road accidents. This project aims to develop a Facial Emotion Recognition (FER) system that monitors the driver's facial expressions to identify emotions and provide instant assistance for safety control. In the initial stage, Viola-Jones face detection was employed to detect the facial region, followed by Butterworth high-pass filtering to enhance the identified region for locating the eye, nose, and mouth regions, using Viola-Jones face detection. Secondly, the Local Binary Patterns (LBP) feature descriptor is utilized to extract features from the identified eye, nose, and mouth regions. Using 3 RGB channels, the extracted features from these three regions are fed into RessNet-50 and EfficientNet deep networks. The outputs of the two deep learning models' classifiers are combined and integrated using two ensemble methods: ensemble maximum voting and ensemble mean. Based on these combining classifier rules, the performance was evaluated on the JAFFE and KMU-FED databases. The experimental results demonstrate that the proposed method can effectively and with higher accuracy than other competitors recognize emotions in the JAFFE and KMU-FED datasets. The novelty and originality of this paper lie in its significant application in the automotive industry. Implementing our proposed method in a system capable of high accuracy and precision can help mitigate numerous driving hazards. Our approach has achieved 99% and 98% accuracy on the JAFFE and KMU-FED databases, respectively. This high level of accuracy, coupled with its practical relevance, underscores the innovative nature of our work. 3504 Image Processing Efficient 3D Shape Matching: Dense Correspondence for non-isometric Deformation Amirfathiyan Amirreza Ebrahimnezhad Hossein Computer Vision Res. Lab., Electrical Engineering Faculty, Sahand University of Technology, Tabriz, Iran. Computer Vision Res. Lab., Electrical Engineering Faculty, Sahand University of Technology, Tabriz, Iran. 1 11 2024 20 4 162 172 19 10 2024 15 12 2024 This paper presents an application of deep learning in computer graphics, utilizing learn-based networks for 3D shape matching. We propose an efficient method for shape matching between 3D models with non-isometric deformation. Our method organizes intrinsic and directional attributes in a structured manner. For this purpose, we use a hybrid feature derived from Diffusion-Net and spectral features. In fact, we combine learned-based intrinsic properties with orientation-preserving features and demonstrate the effectiveness of our method. We achieve this by first extracting features from Diffusion-Net. Then, we compute two maps based on the functional map networks to obtain intrinsic and directional features. Finally, we combine them to achieve a desired map that can resolve symmetry ambiguities on models with high deformation. Quantitative results on the TOSCA dataset indicate that the proposed method achieves lowest average geodetic error of 0.0023, outperforming state-of-the-art methods and reducing the error by 70.66%. We demonstrate that our method outperforms similar approaches by leveraging an accurate feature extractor and effective geometric regularizers, allowing for better handling of non-isometric shapes and resulting in reduced matching errors. 3509 Radar and Sonar Convolutional Neural Network Classification using Three Cepstrums Combinations with Time, Time Derivative and Reassigned STFT of Doppler Signatures from Radar Human Locomotion Yessad Dalila Faculty of Sciences and Technology, Abdelhafid Boussouf University 1 11 2024 20 4 173 182 22 10 2024 29 12 2024 This paper introduces the CTDRCepstrum, a novel feature extraction technique designed to differentiate various human activities using Doppler radar classification. Real data were collected from a Doppler radar system, capturing nine return echoes while monitoring three distinct human activities: walking, fast walking, and running. These activities were performed by three subjects, either individually or in pairs. We focus on analyzing the Doppler signatures using time-frequency reassignment, emphasizing its advantages such as improved component separability. The proposed CTDRCepstrum explores different window functions, transforming each echo signal into three forms of Short-Time Fourier Transform reassignments (RSTFT): time RSTFT (TSTFT), time derivative RSTFT (TDSTFT), and reassigned STFT (RSTFT). A convolutional neural network (CNN) model was then trained using the feature vector, which is generated by combining the cepstral analysis results of each RSTFT form. Experimental results demonstrate the effectiveness of the proposed method, achieving a remarkable classification accuracy of 99.83% by using the Bartlett-Hanning window to extract key features from real-time Doppler radar data of moving targets.