Electronics

20 pages, 2900 KiB

Open AccessArticle

KAN–CNN: A Novel Framework for Electric Vehicle Load Forecasting with Enhanced Engineering Applicability and Simplified Neural Network Tuning

by Zhigang Pei, Zhiyuan Zhang, Jiaming Chen, Weikang Liu, Bailian Chen, Yanping Huang, Haofan Yang and Yijun Lu

Electronics 2025, 14(3), 414; https://doi.org/10.3390/electronics14030414 (registering DOI) - 21 Jan 2025

Abstract

Electric Vehicle (EV) load forecasting is critical for optimizing resource allocation and ensuring the stability of modern energy systems. However, traditional machine learning models, predominantly based on Multi-Layer Perceptrons (MLPs), encounter substantial challenges in modeling the complex, nonlinear, and dynamic patterns inherent in [...] Read more.

Electric Vehicle (EV) load forecasting is critical for optimizing resource allocation and ensuring the stability of modern energy systems. However, traditional machine learning models, predominantly based on Multi-Layer Perceptrons (MLPs), encounter substantial challenges in modeling the complex, nonlinear, and dynamic patterns inherent in EV charging data, often leading to overfitting and high computational costs. To overcome these limitations, this study introduces KAN–CNN, a novel hybrid architecture that integrates Kolmogorov–Arnold Networks (KANs) into traditional machine learning frameworks, specifically Convolutional Neural Networks (CNNs). By combining the spatial feature extraction strength of CNNs with the adaptive nonlinearity of KAN, KAN–CNN achieves superior feature representation and modeling flexibility. The key innovations include bottleneck KAN convolutional layers for reducing parameter complexity, Self-Attention Kolmogorov–Arnold Network with Global Nonlinearity (Self-KAGN) Attention to enhance global dependency modeling, and Focal KAGN Modulation for dynamic feature refinement. Furthermore, regularization techniques such as L1/L2 penalties, dropout, and Gaussian noise injection are utilized to enhance the model’s robustness and generalization capability. When applied to EV load forecasting, KAN–CNN demonstrates prediction accuracy comparable to state-of-the-art methods while significantly reducing computational overhead and simplifying parameter tuning. This work bridges the gap between theoretical innovations and practical applications, offering a robust and efficient solution for dynamic energy system challenges. Full article

(This article belongs to the Special Issue Applications of Data Analytics and Artificial Intelligence in Electric Vehicles)

► Show Figures

Figure 1

11 pages, 2946 KiB

Open AccessArticle

Negative Capacitance Analysis of Multi-Quantum-Well Light-Emitting Diodes

by Yang Xiao, Xiaoyu Feng, Yuan Meng, Longzhen He, Pengzhe Zhang, Dongyan Zhang, Shoushuai Gao, Philip Shields, Haitao Tian and Hongwei Liu

Electronics 2025, 14(3), 413; https://doi.org/10.3390/electronics14030413 (registering DOI) - 21 Jan 2025

Abstract

To explain the negative capacitance (NC) characteristic of multi-quantum-well (MQW) LEDs, we calculated the continuity equation for a 10-period AlGaInP/GaInP multi-quantum-well (MQW) LED with a mesa size of 90 × 150 μm and build an MQW LED capacitor model. The carrier concentrations and [...] Read more.

To explain the negative capacitance (NC) characteristic of multi-quantum-well (MQW) LEDs, we calculated the continuity equation for a 10-period AlGaInP/GaInP multi-quantum-well (MQW) LED with a mesa size of 90 × 150 μm and build an MQW LED capacitor model. The carrier concentrations and capacitance–voltage characteristics across every quantum well region were analyzed by accounting for carrier spontaneous, Auger, and SRH recombination. In our model, a dynamic carrier lifetime iteration method with an iterative error of less than 1 × 10⁻¹⁴ ns was used to decouple carrier lifetime and electric field variables in the carrier continuity equation, providing a new way to enhance the accuracy of MQW continuous equations. Based on the calculated carrier concentration and lifetime of MQWs, we derived a multilayer epitaxial material LED capacitance equivalent circuit. The theoretical model characterizes the negative capacitance phenomenon at 1.75 V, which is consistent with the actual test results of the sample. Our theoretical analysis indicates that the negative capacitance mainly comes from the carrier recombination in the MQW region. Under low-frequency AC bias conditions, the negative capacitance phenomenon becomes more obvious. This work provides a useful reference for analyzing the capacitance and bandwidth characteristics of LEDs in the fields of display dimming and visible-light communication. Full article

(This article belongs to the Section Optoelectronics)

► Show Figures

Figure 1

28 pages, 5356 KiB

Open AccessArticle

Temporal Adaptive Attention Map Guidance for Text-to-Image Diffusion Models

by Sunghoon Jung and Yong Seok Heo

Electronics 2025, 14(3), 412; https://doi.org/10.3390/electronics14030412 (registering DOI) - 21 Jan 2025

Abstract

Text-to-image generation aims to create visually compelling images aligned with input prompts, but challenges such as subject mixing and subject neglect, often caused by semantic leakage during the generation process, remain, particularly in multi-subject scenarios. To mitigate this, existing methods optimize attention maps [...] Read more.

Text-to-image generation aims to create visually compelling images aligned with input prompts, but challenges such as subject mixing and subject neglect, often caused by semantic leakage during the generation process, remain, particularly in multi-subject scenarios. To mitigate this, existing methods optimize attention maps in diffusion models, using static loss functions at each time step, often leading to suboptimal results due to insufficient consideration of varying characteristics across diffusion stages. To address this problem, we propose a novel framework that adaptively guides the attention maps by dividing the diffusion process into four intervals: initial, layout, shape, and refinement. We adaptively optimize attention maps using interval-specific strategies and a dynamic loss function. Additionally, we introduce a seed filtering method based on the self-attention map analysis to detect and address the semantic leakage by restarting the generation process with new noise seeds when necessary. Extensive experiments on various datasets demonstrate that our method achieves significant improvements in generating images aligned with input prompts, outperforming previous approaches both quantitatively and qualitatively. Full article

(This article belongs to the Special Issue Image Fusion and Image Processing)

► Show Figures

Figure 1

25 pages, 15082 KiB

Open AccessArticle

A Sub-6GHz Two-Port Crescent MIMO Array Antenna for 5G Applications

by Heba Ahmed, Allam M. Ameen, Ahmed Magdy, Ahmed Nasser and Mohammed Abo-Zahhad

Electronics 2025, 14(3), 411; https://doi.org/10.3390/electronics14030411 (registering DOI) - 21 Jan 2025

Abstract

The fifth generation of wireless communication (5G) technology is becoming more innovative with the increasing need for high data rates because of the incremental rapidity of mobile data growth. In 5G systems, enhancing device-to-device communication, ultra-low latency (1 ms), outstanding dependability, significant flexibility, [...] Read more.

The fifth generation of wireless communication (5G) technology is becoming more innovative with the increasing need for high data rates because of the incremental rapidity of mobile data growth. In 5G systems, enhancing device-to-device communication, ultra-low latency (1 ms), outstanding dependability, significant flexibility, and data throughput (up to 20 Gbps) is considered one of the most essential factors for wireless networks. To meet these objectives, a sub-6 5G wideband multiple-input multiple-output (MIMO) array microstrip antenna for 5G Worldwide Interoperability for Microwave Access (WiMAX) applications on hotspot devices has been proposed in this research. The 1 × 4 MIMO array radiating element antenna with a partial ground proposed in this research complies with the 5G application standard set out by the Federal Communications Commission. The planned antenna configuration consists of a hollow, regular circular stub patch antenna shaped like a crescent with a rectangular defect at the top of the patch. The suggested structure is mounted on an FR-4 substrate with a thickness “h” of 1.6, a permittivity “

ε_{r}

” of 4.4, and a tangential loss of 0.02. The proposed antenna achieves a high radiation gain and offers a frequency spectrum bandwidth of 3.01 GHz to 6.5 GHz, covering two 5G resonant frequencies “

f_{r}

” of 3.5 and 5.8 GHz as the mid-band, which yields a gain of 7.66 dBi and 7.84 dBi, respectively. MIMO antenna parameters are examined and introduced to assess the system’s performance. Beneficial results are obtained, with the channel capacity loss (CCL) tending to 0.2 bit/s/Hz throughout the operating frequency band, the envelope correlation coefficient (ECC) yielding 0.02, a mean effective gain (MEG) of less than −6 dB over the operating frequency band, and a total active reflection coefficient (TARC) of less than −10 dB; the radiation efficiency is equal to 71.5%, maintaining impedance matching as well as good mutual coupling among the adjacent parameters. The suggested antenna has been implemented and experimentally tested using the 5G system Open Air Interface (OAI) platform, which operates at sub-6 GHz, yielding −67 dBm for the received signal strength indicator (RSSI), and superior frequency stability, precision, and reproducibility for the signal-to-interference-plus-noise ratio (SINR) and a high level of positivity in the power headroom report (PHR) 5G system performance report, confirming its operational effectiveness in 5G WiMAX (Worldwide Interoperability for Microwave Access) application. Full article

(This article belongs to the Section Microwave and Wireless Communications)

► Show Figures

Figure 1

17 pages, 642 KiB

Open AccessFeature PaperArticle

A Distributed Trustable Framework for AI-Aided Anomaly Detection

by Nikolaos Nomikos, George Xylouris, Gerasimos Patsourakis, Vasileios Nikolakakis, Anastasios Giannopoulos, Charilaos Mandilaris, Panagiotis Gkonis, Charalabos Skianis and Panagiotis Trakadas

Electronics 2025, 14(3), 410; https://doi.org/10.3390/electronics14030410 (registering DOI) - 21 Jan 2025

Abstract

The evolution towards sixth-generation (6G) networks requires new architecture enhancements to support the broad device ecosystem, comprising users, machines, autonomous vehicles, and Internet-of-things devices. Moreover, high heterogeneity in the desired quality-of-service (QoS) is expected, as 6G networks will offer extremely low-latency and high-throughput [...] Read more.

The evolution towards sixth-generation (6G) networks requires new architecture enhancements to support the broad device ecosystem, comprising users, machines, autonomous vehicles, and Internet-of-things devices. Moreover, high heterogeneity in the desired quality-of-service (QoS) is expected, as 6G networks will offer extremely low-latency and high-throughput services and error-free communication. This complex environment raises significant challenges in resource management while adhering to security and privacy constraints due to the plethora of data generation endpoints. Considering the advances in AI/ML-aided integration in wireless networks and recent efforts on the network data analytics function (NWDAF) by the 3rd generation partnership project (3GPP), this work presents an AI/ML-aided distributed trustable engine (DTE), collecting data from diverse sources of the 6G infrastructure and deploying ML methods for anomaly detection against diverse threat types. Moreover, we present the DTE architecture and its components, providing data management, AI/ML model training, and classification capabilities for anomaly detection. To promote privacy-aware networking, a federated learning (FL) framework to extend the DTE is discussed. Then, the anomaly detection capabilities of the AI/ML-aided DTE are presented in detail, together with the ML model training process, which considers various ML models. For this purpose, we use two open datasets representing attack scenarios in the core and the edge parts of the network. Experimental results, including an ensemble learning method and different supervised learning alternatives, show that the AI/ML-aided DTE can efficiently train ML models with reduced dimensionality and deploy them in diverse cybersecurity scenarios to improve anomaly detection in 6G networks. Full article

(This article belongs to the Special Issue Recent Advances and Challenges in IoT, Cloud and Edge Coexistence)

► Show Figures

Figure 1

24 pages, 7874 KiB

Open AccessArticle

Power Insulator Defect Detection Method Based on Enhanced YOLOV7 for Aerial Inspection

by Jun Hu, Wenwei Wan, Peng Qiao, Yongqi Zhou and Aiguo Ouyang

Electronics 2025, 14(3), 408; https://doi.org/10.3390/electronics14030408 (registering DOI) - 21 Jan 2025

Abstract

As a principal insulating component in power transmission systems, the integrity of the insulator is of paramount importance for ensuring the safe and reliable operation of transmission lines. While the deployment of aerial photography technology has markedly enhanced the efficacy of power facility [...] Read more.

As a principal insulating component in power transmission systems, the integrity of the insulator is of paramount importance for ensuring the safe and reliable operation of transmission lines. While the deployment of aerial photography technology has markedly enhanced the efficacy of power facility inspections, the intricate backgrounds, multifarious viewpoint alterations, and erratic lighting circumstances inherent in the captured images present novel challenges for the algorithmic detection of insulator defects. To address these issues, this study proposes an enhanced version of the YOLOV7 detection algorithm. The introduction of the contextual transformer network (CoTNet) structure and an EMA attention mechanism enhances the model’s capacity to perceive global contextual information in images and to model long-distance feature dependencies. Experiments based on a real aerial photography dataset demonstrate that the proposed algorithm outperforms the benchmark model in all key performance indicators, including accuracy, recall, and F1 score, which improved by 0.6%, 1.8%, and 0.8%, respectively. Additionally, the average precision (mAP@[0.5]) and mAP@[0.5:0.95] improved by 0.6% and 4.4%, respectively. The superiority of the algorithm in feature extraction and target localization is verified through Grad-CAM visual analysis, which provides a high-precision detection method for intelligent inspection of power transmission systems. Full article

(This article belongs to the Special Issue Image Processing Based on Convolution Neural Network)

► Show Figures

Figure 1

2 pages, 123 KiB

Open AccessEditorial

Smart Electronics, Energy, and IoT Infrastructures for Smart Cities

by Zheng Xu and Jemal Abawajy

Electronics 2025, 14(3), 407; https://doi.org/10.3390/electronics14030407 (registering DOI) - 21 Jan 2025

Abstract

The incorporation of advanced AI into energy and grids necessitates a comprehensive understanding of smart cities [...] Full article

(This article belongs to the Special Issue Smart Electronics, Energy, and IoT Infrastructures for Smart Cities)

14 pages, 2608 KiB

Open AccessArticle

Defense Scheme of Federated Learning Based on GAN

by Qing Zhang, Ping Zhang, Wenlong Lu, Xiaoyu Zhou and An Bao

Electronics 2025, 14(3), 406; https://doi.org/10.3390/electronics14030406 (registering DOI) - 21 Jan 2025

Abstract

Federated learning (FL), as a distributed learning mechanism, can have model training completed without directly uploading original data, effectively reducing the risk of privacy leakage. However, through the shared gradient information, research shows that adversaries may reconstruct the original data. To further protect [...] Read more.

Federated learning (FL), as a distributed learning mechanism, can have model training completed without directly uploading original data, effectively reducing the risk of privacy leakage. However, through the shared gradient information, research shows that adversaries may reconstruct the original data. To further protect the privacy of federated learning, a federated learning defense scheme is proposed based on generative adversarial networks (GAN), which is combined with adaptive differential privacy. Firstly, the real data distribution features are learned through GAN, and replaceable pseudo data are generated. Then, the pseudo data are added with adaptive noise. Finally, the pseudo gradient generated by the pseudo data in the model is used to replace the real gradient so that adversaries cannot obtain the real gradient to further protect the privacy of user data. After simulation experiments are carried out on the MNIST dataset, the algorithm is verified using the gradient attack method. The experimental results show that the proposed algorithm is superior to the federated learning algorithm based on differential privacy in accuracy. Compared with the FedAvg algorithm, only 0.48% accuracy is lost. Therefore, it achieves a good balance between algorithm accuracy and data privacy. Full article

(This article belongs to the Special Issue Security and Privacy for AI)

► Show Figures

Figure 1

17 pages, 396 KiB

Open AccessArticle

An NTRU-Based Key Encapsulation Scheme for Underwater Acoustic Communication

by Peng He and Ming Xu

Electronics 2025, 14(3), 405; https://doi.org/10.3390/electronics14030405 (registering DOI) - 21 Jan 2025

Abstract

With the increasing emphasis on safeguarding maritime sovereignty and developing marine resources, the security of underwater acoustic communication has risen to a new level of importance. Given the complex environmental challenges faced by underwater acoustic channels, this paper proposes an NTRU-based key encapsulation [...] Read more.

With the increasing emphasis on safeguarding maritime sovereignty and developing marine resources, the security of underwater acoustic communication has risen to a new level of importance. Given the complex environmental challenges faced by underwater acoustic channels, this paper proposes an NTRU-based key encapsulation scheme designed to ensure secure and reliable underwater data transmission, while maintaining privacy and integrity. In the public–private key pair generation phase, a ring sampling technique is employed to generate a compact NTRU trapdoor, which not only guarantees security but also effectively reduces the communication overhead. During the encapsulation phase, underwater acoustic channel characteristics during communication are introduced as temporary identity information to ensure the confidentiality and reliability of the key encapsulation mechanism. Furthermore, the traditional key encapsulation mechanism is extended by integrating a digital signature process, where the encapsulated ciphertext is signed. The use of digital signature technology verifies the authenticity and integrity of the transmitted data, ensuring that communication data remain secure and unaltered in complex underwater acoustic environments. Finally, we conduct a rigorous correctness analysis and security proofs, demonstrating that the proposed scheme achieves chosen ciphertext security, while meeting the demands of low bandwidth and limited computational capacity in underwater acoustic communication. Full article

(This article belongs to the Special Issue Security, Privacy, Confidentiality and Trust in Blockchain)

► Show Figures

Figure 1

24 pages, 1924 KiB

Open AccessArticle

Efficient Embedded System for Drowsiness Detection Based on EEG Signals: Features Extraction and Hardware Acceleration

by Aymen Zayed, Emanuel Trabes, Jimmy Tarrillo, Khaled Ben Khalifa and Carlos Valderrama

Electronics 2025, 14(3), 404; https://doi.org/10.3390/electronics14030404 (registering DOI) - 21 Jan 2025

Abstract

Drowsiness detection is crucial for ensuring the safety of individuals engaged in high-risk activities. Numerous studies have explored drowsiness detection techniques based on EEG signals, but these have typically been validated on computers, which limits their portability. In this paper, we introduce the [...] Read more.

Drowsiness detection is crucial for ensuring the safety of individuals engaged in high-risk activities. Numerous studies have explored drowsiness detection techniques based on EEG signals, but these have typically been validated on computers, which limits their portability. In this paper, we introduce the design and implementation of a drowsiness detection technique utilizing EEG signals, executed on a Zynq7020 System on Chip (SoC) as part of a Pynq-Z2 module. This approach is more suitable for portable applications. We have implemented the Discrete Wavelet Transform (DWT) and feature extraction functions as intellectual property (IP) cores, while other functions run on the ARM processor of the Zynq7020. Full article

► Show Figures

Figure 1

25 pages, 5896 KiB

Open AccessArticle

Dynamic Adaptive Artificial Hummingbird Algorithm-Enhanced Deep Learning Framework for Accurate Transmission Line Temperature Prediction

by Xiu Ji, Chengxiang Lu, Beimin Xie, Huanhuan Han and Mingge Li

Electronics 2025, 14(3), 403; https://doi.org/10.3390/electronics14030403 (registering DOI) - 21 Jan 2025

Abstract

As power demand increases and the scale of power grids expands, accurately predicting transmission line temperatures is becoming essential for ensuring the stability and security of power systems. Traditional physical and statistical models struggle with complex multivariate time series, often failing to balance [...] Read more.

As power demand increases and the scale of power grids expands, accurately predicting transmission line temperatures is becoming essential for ensuring the stability and security of power systems. Traditional physical and statistical models struggle with complex multivariate time series, often failing to balance short-term fluctuations with long-term dependencies, and their prediction accuracy and adaptability remain limited. To address these challenges, this paper proposes a deep learning model architecture based on the Dynamic Adaptive Artificial Hummingbird Algorithm (DA-AHA), named the DA-AHA-CNN-LSTM-TPA (DA-AHA-CLT). The model integrates convolutional neural networks (CNNs) for local feature extraction, long short-term memory (LSTM) networks for temporal modeling, and temporal pattern attention mechanisms (TPA) for dynamic feature weighting, while the DA-AHA optimizes hyperparameters to enhance prediction accuracy and stability. The traditional artificial hummingbird algorithm (AHA) is further improved by introducing dynamic step-size adjustment, greedy local search, and grouped parallel search mechanisms to balance global exploration and local exploitation. Our experimental results demonstrate that the DA-AHA-CLT model achieves a coefficient of determination (R²) of 0.987, a root-mean-square error (RMSE) of 0.023, a mean absolute error (MAE) of 0.018, and a median absolute error (MedAE) of 0.011, outperforming traditional models such as CNN-LSTM and LSTM-TPA. These findings confirm that the DA-AHA-CLT model effectively captures the complex dynamic characteristics of transmission line temperatures, offering superior performance and robustness in full-time-step prediction tasks, and highlight its potential for solving challenging multivariate time-series forecasting problems in power systems. Full article

► Show Figures

Figure 1

2 pages, 161 KiB

Open AccessCorrection

Correction: Popoola, G.; Sheppard, J. Investigating and Mitigating the Performance–Fairness Tradeoff via Protected-Category Sampling. Electronics 2024, 13, 3024

by Gideon Popoola and John Sheppard

Electronics 2025, 14(3), 402; https://doi.org/10.3390/electronics14030402 (registering DOI) - 21 Jan 2025

Abstract

In the original publication [...] Full article

34 pages, 359 KiB

Open AccessFeature PaperArticle

Model Checking Using Large Language Models—Evaluation and Future Directions

by Sotiris Batsakis, Ilias Tachmazidis, Matthew Mantle, Nikolaos Papadakis and Grigoris Antoniou

Electronics 2025, 14(2), 401; https://doi.org/10.3390/electronics14020401 - 20 Jan 2025

Abstract

Large language models (LLMs) such as ChatGPT have risen in prominence recently, leading to the need to analyze their strengths and limitations for various tasks. The objective of this work was to evaluate the performance of large language models for model checking, which [...] Read more.

Large language models (LLMs) such as ChatGPT have risen in prominence recently, leading to the need to analyze their strengths and limitations for various tasks. The objective of this work was to evaluate the performance of large language models for model checking, which is used extensively in various critical tasks such as software and hardware verification. A set of problems were proposed as a benchmark in this work and three LLMs (GPT-4, Claude, and Gemini) were evaluated with respect to their ability to solve these problems. The evaluation was conducted by comparing the responses of the three LLMs with the gold standard provided by model checking tools. The results illustrate the limitations of LLMs in these tasks, identifying directions for future research. Specifically, the best overall performance (ratio of problems solved correctly) was 60%, indicating a high probability of reasoning errors by the LLMs, especially when dealing with more complex scenarios requiring many reasoning steps, and the LLMs typically performed better when generating scripts for solving the problems rather than solving them directly. Full article

(This article belongs to the Special Issue Advances in Information, Intelligence, Systems and Applications)

29 pages, 1108 KiB

Open AccessArticle

An Evolutionary Deep Learning Framework for Accurate Remaining Capacity Prediction in Lithium-ion Batteries

by Yang Liu, Liangyu Han, Yuzhu Wang, Jinqi Zhu, Bo Zhang and Jia Guo

Electronics 2025, 14(2), 400; https://doi.org/10.3390/electronics14020400 - 20 Jan 2025

Abstract

Accurate remaining capacity prediction (RCP) of lithium-ion batteries (LIBs) is crucial for ensuring their safety, reliability, and performance, particularly amidst the growing energy crisis and environmental concerns. However, the complex aging processes of LIBs significantly hinder accurate RCP, as traditional prediction methods struggle [...] Read more.

Accurate remaining capacity prediction (RCP) of lithium-ion batteries (LIBs) is crucial for ensuring their safety, reliability, and performance, particularly amidst the growing energy crisis and environmental concerns. However, the complex aging processes of LIBs significantly hinder accurate RCP, as traditional prediction methods struggle to effectively capture nonlinear degradation patterns and long-term dependencies. To tackle these challenges, we introduce an innovative framework that combines evolutionary learning with deep learning for RCP. This framework integrates Temporal Convolutional Networks (TCNs), Bidirectional Gated Recurrent Units (BiGRUs), and an attention mechanism to extract comprehensive time-series features and improve prediction accuracy. Additionally, we introduce a hybrid optimization algorithm that combines the Sparrow Search Algorithm (SSA) with Bayesian Optimization (BO) to enhance the performance of the model. The experimental results validate the superiority of our framework, demonstrating its capability to achieve significantly improved prediction accuracy compared to existing methods. This study provides researchers in battery management systems, electric vehicles, and renewable energy storage with a reliable tool for optimizing lithium-ion battery performance, enhancing system reliability, and addressing the challenges of the new energy industry. Full article

(This article belongs to the Special Issue Advanced Machine Learning, Pattern Recognition, and Deep Learning Technologies: Methodologies and Applications)

19 pages, 2356 KiB

Open AccessArticle

Optimizing 3D Point Cloud Reconstruction Through Integrating Deep Learning and Clustering Models

by Seyyedbehrad Emadi and Marco Limongiello

Electronics 2025, 14(2), 399; https://doi.org/10.3390/electronics14020399 - 20 Jan 2025

Abstract

Noise in 3D photogrammetric point clouds—both close-range and UAV-generated—poses a significant challenge to the accuracy and usability of digital models. This study presents a novel deep learning-based approach to improve the quality of point clouds by addressing this issue. We propose a two-step [...] Read more.

Noise in 3D photogrammetric point clouds—both close-range and UAV-generated—poses a significant challenge to the accuracy and usability of digital models. This study presents a novel deep learning-based approach to improve the quality of point clouds by addressing this issue. We propose a two-step methodology: first, a variational autoencoder reduces features, followed by clustering models to assess and mitigate noise in the point clouds. This study evaluates four clustering methods—k-means, agglomerative clustering, Spectral clustering, and Gaussian mixture model—based on photogrammetric parameters, reprojection error, projection accuracy, angles of intersection, distance, and the number of cameras used in tie point calculations. The approach is validated using point cloud data from the Temple of Neptune in Paestum, Italy. The results show that the proposed method significantly improves 3D reconstruction quality, with k-means outperforming other clustering techniques based on three evaluation metrics. This method offers superior versatility and performance compared to traditional and machine learning techniques, demonstrating its potential to enhance UAV-based surveying and inspection practices. Full article

(This article belongs to the Special Issue Point Cloud Data Processing and Applications)

16 pages, 12452 KiB

Open AccessArticle

Scaling Nuclear Magnetic Resonance with Integrated Planar Coil and Transceiver Front-End: Co-Design Considerations

by Natachai Terawatsakul, Alireza Saberkari, Yuttapoom Puttisong and Morgan Madec

Electronics 2025, 14(2), 398; https://doi.org/10.3390/electronics14020398 - 20 Jan 2025

Abstract

A comprehensive framework for designing a micro-nuclear magnetic resonance (NMR) front-end is presented. Key radio frequency (RF) engineering principles are established to enable efficient excitation and detection of NMR signals. This foundation aims to guide the optimal design of novel handheld NMR devices [...] Read more.

A comprehensive framework for designing a micro-nuclear magnetic resonance (NMR) front-end is presented. Key radio frequency (RF) engineering principles are established to enable efficient excitation and detection of NMR signals. This foundation aims to guide the optimal design of novel handheld NMR devices operating with magnetic fields (

B_{0}

) below 0.5 Tesla and RF frequencies under 30 MHz. To address the complexities of signal-to-noise ratio optimization in this regime, a specialized metric called the coil performance factor (CPF) is introduced, emphasizing the role of coil design. Through systematic optimization under realistic constraints, an optimal coil configuration maximizing the CPF is identified. This design, with three turns, a coil width of 0.22 mm, and a coil spacing of 0.15 mm, achieves an optimal balance between magnetic field strength, homogeneity, and noise. This work serves as a valuable resource for engineers developing optimized coil designs and RF solutions for handheld NMR devices, providing clear explanations of essential concepts and a practical design methodology. Full article

(This article belongs to the Special Issue RF/MM-Wave Circuits Design and Applications, 2nd Edition)

► Show Figures

Figure 1

24 pages, 7528 KiB

Open AccessArticle

EOS: Edge-Based Operation Skip Scheme for Real-Time Object Detection Using Viola-Jones Classifier

by Cheol-Ho Choi, Joonhwan Han, Hyun Woo Oh, Jeongwoo Cha and Jungho Shin

Electronics 2025, 14(2), 397; https://doi.org/10.3390/electronics14020397 - 20 Jan 2025

Abstract

Machine learning-based object detection systems are preferred due to their cost-effectiveness compared to deep learning approaches. Among machine learning methods, the Viola-Jones classifier stands out for its reasonable accuracy and efficient resource utilization. However, as the number of classification iterations increases or the [...] Read more.

Machine learning-based object detection systems are preferred due to their cost-effectiveness compared to deep learning approaches. Among machine learning methods, the Viola-Jones classifier stands out for its reasonable accuracy and efficient resource utilization. However, as the number of classification iterations increases or the resolution of the input image increases, the detection processing speed may decrease. To address the detection speed issue related to input image resolution, an improved edge component calibration method is applied. Additionally, an edge-based operation skip scheme is proposed to overcome the detection processing speed problem caused by the number of classification iterations. Our experiments using the FDDB public dataset show that our method reduces classification iterations by 24.6157% to 84.1288% compared to conventional methods, except for our previous study. Importantly, our method maintains detection accuracy while reducing classification iterations. This result implies that our method can realize almost real-time object detection when implemented on field-programmable gate arrays. Full article

(This article belongs to the Special Issue Artificial Intelligence in Image and Video Processing)

► Show Figures

Figure 1

16 pages, 1161 KiB

Open AccessArticle

Multiplex Graph Contrastive Learning with Soft Negatives

by Zhenhao Zhao, Minhong Zhu, Chen Wang, Sijia Wang, Jiqiang Zhang, Li Chen and Weiran Cai

Electronics 2025, 14(2), 396; https://doi.org/10.3390/electronics14020396 - 20 Jan 2025

Abstract

Graph Contrastive Learning (GCL) seeks to learn nodal or graph representations that contain maximal consistent information from graph-structured data. While node-level contrasting modes are dominating, some efforts have commenced to explore consistency across different scales. Yet, they tend to lose consistent information and [...] Read more.

Graph Contrastive Learning (GCL) seeks to learn nodal or graph representations that contain maximal consistent information from graph-structured data. While node-level contrasting modes are dominating, some efforts have commenced to explore consistency across different scales. Yet, they tend to lose consistent information and be contaminated by disturbing features. We propose MUX-GCL, a novel cross-scale contrastive learning framework that addresses these key challenges in GCL by leveraging multiplex representations as effective patches to enhance information consistency. Our method introduces a soft-negative contrasting strategy based on positional affinities to reduce false negatives, thereby minimizing information loss during multi-scale contrasts. While this learning mode minimizes contaminating noises, a commensurate contrasting strategy using positional affinities further avoids information loss by correcting false negative pairs across scales. Extensive downstream experiments demonstrate that MUX-GCL yields multiple state-of-the-art results on public datasets. Our theoretical analysis further guarantees the new objective function as a stricter lower bound of mutual information of raw input features and output embeddings, which rationalizes this paradigm. Full article

► Show Figures

Figure 1

26 pages, 23657 KiB

Open AccessArticle

A Digital Twin Approach for Soil Moisture Measurement with Physically Based Rendering Simulations and Machine Learning

by Ismail Parewai and Mario Köppen

Electronics 2025, 14(2), 395; https://doi.org/10.3390/electronics14020395 - 20 Jan 2025

Abstract

Soil is one of the most important factors of agricultural productivity, directly influencing crop growth, water management, and overall yield. However, inefficient soil moisture monitoring methods, such as manual observation and gravimetric in rural areas, often lead to overwatering or underwatering, wasting resources [...] Read more.

Soil is one of the most important factors of agricultural productivity, directly influencing crop growth, water management, and overall yield. However, inefficient soil moisture monitoring methods, such as manual observation and gravimetric in rural areas, often lead to overwatering or underwatering, wasting resources and reduced yields, and harming soil health. This study offers a digital twin approach for soil moisture measurement, integrating real-time physical data, virtual simulations, and machine learning to classify soil moisture conditions. The digital twin is proposed as a virtual representation of physical soil designed to replicate real-world behavior. We used a multispectral rotocam, and high-resolution soil images were captured under controlled conditions. Physically based rendering (PBR) materials were created from these data and implemented in a game engine to simulate soil properties accurately. Image processing techniques were applied to extract key features, followed by machine learning algorithms to classify soil moisture levels (wet, normal, dry). Our results demonstrate that the Soil Digital Twin replicates real-world behavior, with the Random Forest model achieving a high classification accuracy of 96.66% compared to actual soil. This data-driven approach conveys the potential of the Soil Digital Twin to enhance precision farming initiatives and water use efficiency for sustainable agriculture. Full article

(This article belongs to the Special Issue Pattern Recognition and Image Processing: Latest Advances and Prospects)

► Show Figures

Figure 1

Journal Description

Electronics

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI