Applied Sciences

22 pages, 11126 KiB

Open AccessArticle

Analytical Investigation of Vertical Force Control in In-Wheel Motors for Enhanced Ride Comfort

by Chanoknan Bunlapyanan, Sunhapos Chantranuwathana and Gridsada Phanomchoeng

Appl. Sci. 2024, 14(15), 6582; https://doi.org/10.3390/app14156582 (registering DOI) - 27 Jul 2024

This study explores the effectiveness of vertical force control in in-wheel motors (IWMs) to enhance ride comfort in electric vehicles (EVs). A dynamic vehicle model and a proportional ride-blending controller were used to reduce vertical vibrations of the sprung mass. By converting the [...] Read more.

This study explores the effectiveness of vertical force control in in-wheel motors (IWMs) to enhance ride comfort in electric vehicles (EVs). A dynamic vehicle model and a proportional ride-blending controller were used to reduce vertical vibrations of the sprung mass. By converting the state-space model into a transfer function, the system’s frequency response was evaluated using road profiles generated according to ISO 8608 standards and converted into Power Spectral Density (PSD) inputs. The frequency-weighted acceleration (

a_{w}

) was calculated based on ISO 2631 standards to measure ride comfort improvements. The results showed that increasing the proportional gain (

K_{p}

) effectively reduced the frequency-weighted acceleration and the RMS of the vertical acceleration of the sprung mass. However, the proportional gain could not be increased indefinitely due to the torque limitations of the IWMs. Optimal proportional gains for various road profiles demonstrated significant improvements in ride comfort. This study concludes that advanced suspension technologies, including the proportional ride-blending controller, can effectively mitigate the challenges of increased unsprung mass in IWM vehicles, thereby enhancing ride quality and vehicle dynamics. Full article

(This article belongs to the Special Issue Advances in Vehicle System Dynamics and Control)

► Show Figures

Figure 1

27 pages, 782 KiB

Open AccessArticle

Evaluating Intelligent CPTED Systems to Support Crime Prevention Decision-Making in Municipal Control Centers

by Woochul Choi, Joonyeop Na and Sangkyeong Lee

Appl. Sci. 2024, 14(15), 6581; https://doi.org/10.3390/app14156581 (registering DOI) - 27 Jul 2024

Abstract

To maximize its synergetic effect across the cycle from prevention to response to post-crime management, crime prevention requires a balanced combination of spatial urban design and advanced crime prevention technologies for crime prediction and real-time response. This study derived intelligent Crime Prevention Through [...] Read more.

To maximize its synergetic effect across the cycle from prevention to response to post-crime management, crime prevention requires a balanced combination of spatial urban design and advanced crime prevention technologies for crime prediction and real-time response. This study derived intelligent Crime Prevention Through Environmental Design (CPTED) services and suggested a decision model based on the fuzzy Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to implement these services in municipal control centers. The analysis results are summarized as follows. First, this study established a fuzzy TOPSIS-based decision-making support model enabling local government control centers to effectively select intelligent CPTED service elements. Second, overall, operator-led Closed-Circuit Television (CCTV) and platform control technologies were identified as significant components of intelligent CPTED service elements. Third, a comparison by city size revealed that large cities in the Seoul metropolitan area rated system services for control based on advanced crime prevention infrastructure (e.g., the crime monitoring systems and real-time control drones/robots) relatively higher. In contrast, small and medium-sized cities in other provinces rated services that were perceptible to residents and improved crime-prone environments (e.g., artificial intelligence (AI) video analysis for living safety) relatively higher. Full article

12 pages, 3248 KiB

Open AccessArticle

Timing Performance Testing and Regularity Analysis of eLoran System

by Jianchen Di, Jun Fu, Bing Xu, Miao Wu, Liang Liu and Xianzhou Jin

Appl. Sci. 2024, 14(15), 6580; https://doi.org/10.3390/app14156580 (registering DOI) - 27 Jul 2024

Abstract

Accurate time synchronisation is critical in modern communication, navigation, and scientific research. In this context, the eLoran receiver, which is an advanced timing device, has attracted increasing scholarly attention. This study aims to comprehensively analyse the performance and potential applications of eLoran systems [...] Read more.

Accurate time synchronisation is critical in modern communication, navigation, and scientific research. In this context, the eLoran receiver, which is an advanced timing device, has attracted increasing scholarly attention. This study aims to comprehensively analyse the performance and potential applications of eLoran systems for timing monitoring and to specifically explore the relevant indices of two eLoran receivers. To this end, we evaluated the performance of these receivers in receiving time signals through both simulated and empirical data and conducted a regularity analysis to uncover their potential value in practical applications. The findings demonstrate that the eLoran receiver excels at timing monitoring and provides highly accurate time information. An analysis of the timing performance of the eLoran system improved its accuracy and integrity. Full article

(This article belongs to the Section Electrical, Electronics and Communications Engineering)

► Show Figures

Figure 1

19 pages, 3637 KiB

Open AccessArticle

Semantic Kansei Engineering Approach for Game Controllers and Design Improvement

by A. G. González, O. López, M. M. Pizarro and J. F. Vázquez

Appl. Sci. 2024, 14(15), 6579; https://doi.org/10.3390/app14156579 (registering DOI) - 27 Jul 2024

Abstract

The entertainment industry undergoes constant changes, and video games, along with the ways people enjoy them, evolve dynamically. In this context, there arises a need to innovate in areas such as user–game interaction and the design of conventional controllers to keep up with [...] Read more.

The entertainment industry undergoes constant changes, and video games, along with the ways people enjoy them, evolve dynamically. In this context, there arises a need to innovate in areas such as user–game interaction and the design of conventional controllers to keep up with these transformations. This study has found that the design requirements of the external geometry of video game controllers significantly impact users’ purchasing decisions, who often rely on their sensations and emotions. The primary aim of this research has been to demonstrate that it is possible to develop a specific methodology to establish design guidelines that evaluate geometry and ergonomic aspects, considering users’ emotional needs. To achieve this goal, the Kansei methodology has been employed, an engineering discipline that focuses on product design based on emotional interaction and user sensations when using them. Through advanced statistical techniques such as principal component analysis and linear regression, the set of observations has been reduced and a robust predictive model has been developed. Following this methodological approach, surveys and field studies have been conducted, thus differentiating the project from a conventional engineering approach in various stages. Special attention has been paid to the tactile part of the controller, especially its casing. After thorough analysis, the redesigned casing, incorporating improvements derived from statistical analyses, has been compared to existing market models. The results obtained show a significant improvement compared to existing market models, particularly highlighting enhancements in functional characteristics and ergonomic design, enabling comfortable use for individuals with different hand sizes and preventing the grip surface from being slippery. In summary, this work supports the effectiveness of the Kansei methodology as an approach to adapting product design to users’ perceptions and emotional needs. Full article

(This article belongs to the Special Issue Advances in Manufacturing Ergonomics)

18 pages, 1992 KiB

Open AccessArticle

Optimizing Recurrent Neural Networks: A Study on Gradient Normalization of Weights for Enhanced Training Efficiency

by Xinyi Wu, Bingjie Xiang, Huaizheng Lu, Chaopeng Li, Xingwang Huang and Weifang Huang

Appl. Sci. 2024, 14(15), 6578; https://doi.org/10.3390/app14156578 (registering DOI) - 27 Jul 2024

Abstract

Recurrent Neural Networks (RNNs) are classical models for processing sequential data, demonstrating excellent performance in tasks such as natural language processing and time series prediction. However, during the training of RNNs, the issues of vanishing and exploding gradients often arise, significantly impacting the [...] Read more.

Recurrent Neural Networks (RNNs) are classical models for processing sequential data, demonstrating excellent performance in tasks such as natural language processing and time series prediction. However, during the training of RNNs, the issues of vanishing and exploding gradients often arise, significantly impacting the model’s performance and efficiency. In this paper, we investigate why RNNs are more prone to gradient problems compared to other common sequential networks. To address this issue and enhance network performance, we propose a method for gradient normalization of network weights. This method suppresses the occurrence of gradient problems by altering the statistical properties of RNN weights, thereby improving training effectiveness. Additionally, we analyze the impact of weight gradient normalization on the probability-distribution characteristics of model weights and validate the sensitivity of this method to hyperparameters such as learning rate. The experimental results demonstrate that gradient normalization enhances the stability of model training and reduces the frequency of gradient issues. On the Penn Treebank dataset, this method achieves a perplexity level of 110.89, representing an 11.48% improvement over conventional gradient descent methods. For prediction lengths of 24 and 96 on the ETTm1 dataset, Mean Absolute Error (MAE) values of 0.778 and 0.592 are attained, respectively, resulting in 3.00% and 6.77% improvement over conventional gradient descent methods. Moreover, selected subsets of the UCR dataset show an increase in accuracy ranging from 0.4% to 6.0%. The gradient normalization method enhances the ability of RNNs to learn from sequential and causal data, thereby holding significant implications for optimizing the training effectiveness of RNN-based models. Full article

(This article belongs to the Section Computing and Artificial Intelligence)

21 pages, 9916 KiB

Open AccessArticle

Reservoir Characteristics of Marine–Continental Transitional Taiyuan Formation Shale and Its Influence on Methane Adsorption Capacity: A Case Study in Southern North China Basin

by Wei Jiang and Yang Hu

Appl. Sci. 2024, 14(15), 6577; https://doi.org/10.3390/app14156577 (registering DOI) - 27 Jul 2024

Abstract

To further study the reservoir characteristics and adsorption capacity of the Taiyuan Formation shale in the South North China Basin (SNCB), the pore structure and adsorption capacity of shale are discussed using various analysis tests, including elemental geochemistry, organic geochemistry, mineral composition, low-temperature [...] Read more.

To further study the reservoir characteristics and adsorption capacity of the Taiyuan Formation shale in the South North China Basin (SNCB), the pore structure and adsorption capacity of shale are discussed using various analysis tests, including elemental geochemistry, organic geochemistry, mineral composition, low-temperature nitrogen adsorption (LTNA), and methane adsorption experiments. The results indicate that the Taiyuan Formation shale formed in a poor oxygen and anaerobic sedimentary environment in still water. The average value of total organic carbon (TOC) content is 2.37%. The organic matter type mainly consists of type III kerogen. The vitinite reflectance (Ro) ranges from 3.11% to 3.50%. The clay mineral content varies greatly, averaging at 40.7%, while the quartz content averages at 37.7%. The Taiyuan Formation shale mainly develops interparticle (InterP) pores, followed by organic pores, intraparticle (IntraP) pores, solution pores, and microfractures. BET specific surface area (SSA) is between 9.47 m²/g and 22.14 m²/g, while pore volume (PV) ranges from 0.0098 cm³/g to 0.022 cm³/g, indicating favorable conditions for shale gas storage. According to the results of the CH₄ adsorption experiment, Langmuir volume from Taiyuan Formation shales exhibits 1.35~4.30 cm³/g, indicating excellent adsorption capacity. TOC content shows a positive correlation with both Langmuir volume and BET SSA from Taiyuan Formation shales, suggesting that TOC plays a crucial role in controlling microscopic pores and gas adsorption capacity. Organic matter enhances the shale adsorption capacity by providing abundant pore SSA. Due to formation compaction, the pore size of clay minerals decreases, leading to an increase in pore SSA, while kaolinite exhibits weak hydrophilic ability. Consequently, with the increase in clay minerals and kaolinite content, the shale adsorption capacity is enhanced to a certain extent. However, an increase in the carbonate mineral content may result in a decrease in the proportion of clay minerals, therefore reducing the CH₄ adsorption capacity of shale. Full article

(This article belongs to the Special Issue Advances in Unconventional Natural Gas: Exploration and Development)

14 pages, 4082 KiB

Open AccessArticle

Influences of Different Selective Laser Melting Machines on the Microstructures and Mechanical Properties of Co–Cr–Mo Alloys

by Atsushi Takaichi, Yuka Kajima, Hein Linn Htat and Noriyuki Wakabayashi

Appl. Sci. 2024, 14(15), 6576; https://doi.org/10.3390/app14156576 (registering DOI) - 27 Jul 2024

Abstract

Dental prostheses have been fabricated using various selective laser melting (SLM) machines; however, the impact of the type of machine on the microstructure and mechanical properties of Co–Cr–Mo alloys remains unclear. In this study, we prepared samples using two SLM machines (the small [...] Read more.

Dental prostheses have been fabricated using various selective laser melting (SLM) machines; however, the impact of the type of machine on the microstructure and mechanical properties of Co–Cr–Mo alloys remains unclear. In this study, we prepared samples using two SLM machines (the small M100 and mid-sized M290) with different beam spot sizes (40 and 100 µm, respectively). The microstructures and tensile properties of the heated (1150 °C for 60 min) and as-built samples were evaluated. The grain sizes of the M100 samples were smaller than those of the M290 samples due to the small beam spot size of the M100 machine. Both heated samples exhibited recrystallized equiaxed grains; however, the amount of non-recrystallized grains remaining in the M290 sample exceeded that in the M100 sample. This suggests that the M100 samples recrystallized faster than the M290 samples after heating. The elongation of the M100 samples was higher than that of the M290 samples in the as-built and heated states, owing to the smaller grain size of the M100 samples. A comparison of the M100 and M290 SLM machines indicated that the M100 was suitable for producing dental prostheses owing to its good elongation and rapid recrystallization features, which shorten its post-heat-treatment duration. Full article

(This article belongs to the Special Issue Three-Dimensional (3D) Printing in Dentistry: Technologies, Influencing Factors and Applications)

► Show Figures

Figure 1

10 pages, 310 KiB

Open AccessArticle

Sport and Physical Activity Participation by Weight Groups in School-Aged Hungarian Children

by Martina Uvacsek, Bernadett Wágner and Georgina Simkó

Appl. Sci. 2024, 14(15), 6575; https://doi.org/10.3390/app14156575 (registering DOI) - 27 Jul 2024

Abstract

The aim of this study was to compare physical activity and organized sport participation survey data between different weight classifications in children. In the cross-sectional online data collection, 677 parents (42.8 ± 6.4 years old) provided information about 677 children (10.9 ± 2.5 [...] Read more.

The aim of this study was to compare physical activity and organized sport participation survey data between different weight classifications in children. In the cross-sectional online data collection, 677 parents (42.8 ± 6.4 years old) provided information about 677 children (10.9 ± 2.5 years old). On average, 77% of the children achieved 60 min of MVPA/day and 63% participated in a club or organized sport. Most of them (63%) were in the healthy weight category, 14% of the children were underweight, and 22% of the children were overweight or obese. The parental support for physical activity was 83%. Club or organized sport participation (OR = 0.56 CI: 0.34–0.91; p < 0.02) and a parent with a higher educational level (OR = 0.25 CI:0.14–0.43; p < 0.01) were found to be preventive for obesity. Football was the most popular and highly chosen sport activity. Regarding sport activities with parents, cycling, walking, and football were the most popular, independently of the weight classification of the child. The overweight and obese children had almost the same participation rate in sport activities than others. Increased physical activity might be explained by the daily physical education in schools and governmental support. Further studies and measures are needed to prevent overweight and obesity in children of parents with a lower educational level. Full article

(This article belongs to the Special Issue Human Performance and Health in Sport and Exercise)

22 pages, 4095 KiB

Open AccessArticle

Estimation of Photovoltaic Potential of Solar-Powered Electric Vehicle: Case Study of Commuters on Donghae Expressway, Korea

by Jangwon Suh

Appl. Sci. 2024, 14(15), 6574; https://doi.org/10.3390/app14156574 (registering DOI) - 27 Jul 2024

Abstract

Studies on solar electric vehicles (EVs) have focused on calculating the power generation in a specific environment without discussing its practical utility. To expand the awareness of the utility of solar EVs, their potential should be evaluated by considering the operation methods of [...] Read more.

Studies on solar electric vehicles (EVs) have focused on calculating the power generation in a specific environment without discussing its practical utility. To expand the awareness of the utility of solar EVs, their potential should be evaluated by considering the operation methods of users. This study investigated the photovoltaic (PV) potential of an EV integrated with PV modules while driving on an expressway. Tunnel and shadow areas were identified to determine unpowered areas on the expressway. The PVWatts model was used to evaluate the PV potential by the time of the year. For a single vehicle traveling at 60 km/h on the Donghae expressway section during both the summer and winter solstices, the amount of power generation is within 0.208–0.317 kWh, corresponding to 0.94–1.43% of the electricity consumed for driving. Furthermore, this study assumed that office workers commute on the Donghae expressway. Under the scenario considering the time of operation (traveling to and from work and parking at work) and the shading ratio, the rechargeable amount was more than 10% of the electricity consumption. The results showed that solar roofs are potential charging supplements for EV batteries. This study can provide the efficacy and optimal operation method of solar EVs for commuters. Full article

(This article belongs to the Special Issue Renewable Energy for Sustainable Transports: The Potential of the Electricity and Hydrogen)

► Show Figures

Figure 1

24 pages, 896 KiB

Open AccessArticle

Comparative Analysis of SAAS Model and NPC Integration for Enhancing VR Shopping Experiences

by Surasachai Doungtap, Jenq-Haur Wang and Varinya Phanichraksaphong

Appl. Sci. 2024, 14(15), 6573; https://doi.org/10.3390/app14156573 (registering DOI) - 27 Jul 2024

Abstract

This article examines the incorporation of the Shopping Assistance Automatic Suggestion (SAAS) model into Virtual Reality (VR) environments in order to improve the online shopping experience. The SAAS model employs sophisticated deep learning methods to offer customized product recommendations, which are conveyed by [...] Read more.

This article examines the incorporation of the Shopping Assistance Automatic Suggestion (SAAS) model into Virtual Reality (VR) environments in order to improve the online shopping experience. The SAAS model employs sophisticated deep learning methods to offer customized product recommendations, which are conveyed by non-player characters (NPCs) via voice-based interactions. Our goal is to develop an interactive shopping experience that replicates real-life interactions by integrating AI-powered recommendations with immersive VR technology. We gather and standardize data from several open commerce databases, such as Amazon Product and Customer Reviews. The SAAS model, in conjunction with GPT-3, BERT, and T5, undergoes training and testing to evaluate its effectiveness across multiple criteria. The results demonstrate that the SAAS model surpasses other models in delivering contextually aware and pertinent recommendations. The integration process outlines the specific steps involved in capturing, processing, and transforming user interactions in virtual reality (VR) into vocal suggestions provided by non-player characters (NPCs). This strategy improves customization and utilizes the immersive features of virtual reality to effectively engage people. The results of our research establish a higher standard for e-commerce, with the goal of enhancing the user experience of online purchasing by making it more instinctive, engaging, and pleasurable. Full article

(This article belongs to the Special Issue Enhancing User Experience in Virtual Reality Environments: Innovative Interaction Design Strategies)

11 pages, 3676 KiB

Open AccessBrief Report

Plastome Evolution of Asyneuma japonicum: Insights into Structural Variation, Genomic Divergence, and Phylogenetic Tree

by Byeong-Seon Park, Won-Sub Yoon, Chang-Kug Kim and Yong-Kab Kim

Appl. Sci. 2024, 14(15), 6572; https://doi.org/10.3390/app14156572 (registering DOI) - 27 Jul 2024

Abstract

Asyneuma japonicum is an ornamental flowering plant in East Asia. The genus Asyneuma is difficult to distinguish taxonomically because of its morphological similarities with the genus Campanula. We constructed the first complete plastome of A. japonicum (NCBI accession number: OR805474) using the Illumina [...] Read more.

Asyneuma japonicum is an ornamental flowering plant in East Asia. The genus Asyneuma is difficult to distinguish taxonomically because of its morphological similarities with the genus Campanula. We constructed the first complete plastome of A. japonicum (NCBI accession number: OR805474) using the Illumina platform. This plastome is a circular ring structure with a length of 185,875 base pairs. It is organized into four parts: a pair of inverted repeats (33,084 bp each) as well as large (83,795 bp) and small (35,912 bp) single-copy regions. One hundred nine unique genes were encoded in the assembled plastome. Using structural variations, junction boundaries, rearrangements, divergent hotspots, and phylogenetic analysis, we revealed that A. japonicum was in the closest evolutionary position to Hanabusaya asiatica and it had a large evolutionary divergence from the Campanulaceae family due to gene rearrangements. Full article

(This article belongs to the Section Applied Biosciences and Bioengineering)

21 pages, 2108 KiB

Open AccessArticle

Background-Filtering Feature-Enhanced Graph Neural Networks for Few-Shot Learning

by Binbin Wang, Yuemao Wang and Yaoqun Xu

Appl. Sci. 2024, 14(15), 6571; https://doi.org/10.3390/app14156571 (registering DOI) - 27 Jul 2024

Abstract

The fundamental idea behind few-shot learning is to employ sparse labeled data to effectively handle novel tasks, whereas most existing mainstream approaches mostly rely on prior experience gained from previous situations. Nonetheless, effective knowledge transfer is sometimes hampered by constraints on new samples [...] Read more.

The fundamental idea behind few-shot learning is to employ sparse labeled data to effectively handle novel tasks, whereas most existing mainstream approaches mostly rely on prior experience gained from previous situations. Nonetheless, effective knowledge transfer is sometimes hampered by constraints on new samples and barriers between classes. To address these issues, this research presents a novel background-filtering feature-enhanced graph network (BFFE-GNN) that attempts to generate relationships between graphs in order to explicitly describe and transmit inter-class relationships. To specifically address the issue of inadequate information utilization brought on by sample background interference, which is frequent in classification tasks, we employ a novel background-filtering feature-enhanced graph network. Effective data information extraction from complicated datasets is challenging due to the original module’s relatively simple network structure during the feature extraction stage and the interference of the image background. The background-filtering module was specifically introduced, which enhances spatial attention. This not only improves the quality of feature extraction but also effectively lessens the influence of picture background on classification results. In addition, we have improved the background-filtering-based feature gap calculation by implementing a feature-enhanced module. To demonstrate the adaptability of the BFFE-GNN model, we not only ran experiments on two publicly available datasets, MiniImagenet and TiredImagenet, but we also created our own Tool dataset. The method’s exceptional performance and universal applicability in the field of few-shot picture classification are clearly demonstrated by the experimental findings, which indicate that it greatly outperforms the majority of existing similar methods. This discovery establishes a strong basis for further research in the field of few-shot learning in addition to offering fresh insights into the subject. Full article

18 pages, 6692 KiB

Open AccessArticle

Numerically Investigating the Effect of Trim on the Resistance of a Container Ship in Confined and Shallow Water

by Ivana Martić, Barbara Anušić, Nastia Degiuli and Carlo Giorgio Grlj

Appl. Sci. 2024, 14(15), 6570; https://doi.org/10.3390/app14156570 (registering DOI) - 27 Jul 2024

Abstract

One of the promising strategies for the improvement of the resistance characteristics and energy efficiency of a ship is trim optimization. Most of the studies conducted so far regarding trim optimization deal with unrestricted water. The effect of trim on the total resistance [...] Read more.

One of the promising strategies for the improvement of the resistance characteristics and energy efficiency of a ship is trim optimization. Most of the studies conducted so far regarding trim optimization deal with unrestricted water. The effect of trim on the total resistance and its components for the KRISO Container Ship model in confined and shallow water is investigated using computational fluid dynamics. Numerical simulations of resistance tests with and without the free surface effects using two turbulence models are carried out for even keel and four trims in restricted water. A verification study is conducted for the total resistance, sinkage, and trim in terms of grid resolutions and time steps to assess the numerical uncertainty. The numerical results are validated against the experimental ones available in the literature. Performing the double body simulations enabled the analysis of the effect of trim on the resistance components. The numerical results pointed out that by adjusting the trim a reduction in the total resistance in confined and shallow water can be obtained. Full article

(This article belongs to the Section Marine Science and Engineering)

► Show Figures

Figure 1

12 pages, 4462 KiB

Open AccessArticle

Research on the Use of Reflective Thermal Insulation Coating on Railway Tracks and Wagons in Slovak Republic

by Eva Brumercikova, Werner Hofer, Frantisek Brumercik, Bibiana Bukova and Vladislav Zitricky

Appl. Sci. 2024, 14(15), 6569; https://doi.org/10.3390/app14156569 (registering DOI) - 27 Jul 2024

Abstract

Global warming is a process in which the temperature of the entire planet gradually increases. High temperatures affect the surface of the material and cause undesirable effects such as damage because of overheating of the material. This article describes the use of reflective [...] Read more.

Global warming is a process in which the temperature of the entire planet gradually increases. High temperatures affect the surface of the material and cause undesirable effects such as damage because of overheating of the material. This article describes the use of reflective thermal insulation coatings on railway tracks abroad. Part of the article describes the method of applying reflective coatings to railway tracks. The article examines, analyzes, and evaluates the effect of a reflective thermal insulation coating on the railway track and railway wagon in the conditions of the Slovak Republic. Full article

► Show Figures

Figure 1

17 pages, 25328 KiB

Open AccessArticle

Comparison of Residual Network and Other Classical Models for Classification of Interlayer Distresses in Pavement

by Wenlong Cai, Mingjie Li, Guanglai Jin, Qilin Liu and Congde Lu

Appl. Sci. 2024, 14(15), 6568; https://doi.org/10.3390/app14156568 (registering DOI) - 27 Jul 2024

Abstract

Many automatic classification methods published can identify the main hidden distress types of highways, but they cannot meet the precise needs of operation and maintenance. The classification of interlayer distresses based on ground penetrating radar (GPR) images is very important to improve maintenance [...] Read more.

Many automatic classification methods published can identify the main hidden distress types of highways, but they cannot meet the precise needs of operation and maintenance. The classification of interlayer distresses based on ground penetrating radar (GPR) images is very important to improve maintenance efficiency and reduce cost. However, among models of different complexities, which models are suitable for the interlayer distress data needs further verification. Firstly, to cover enough of the variable range of distress samples, the interlayer distress dataset collected containing 32,038 samples was subcategorized into three types: interlayer debonding, interlayer water seepage, and interlayer loosening. Secondly, residual networks (ResNets) that render easier to build shallower or deeper networks (ResNet-4, ResNet-6, ResNet-8, ResNet-10, ResNet-14, ResNet-18, ResNet-34, and ResNet-50) and five classical network models (DenseNet-121, EfficientNet B0, SqueezeNet1_0, MobileNet V2, and VGG-19) were evaluated by training and validation loss, test accuracy, and model complexity. The experimental results show that all models have high test accuracy with little difference, but ResNet-4, ResNet-6, SqueezeNet1_0, and ResNet-8 exhibit no overfitting which means they have good generalization performance. Full article

(This article belongs to the Special Issue Damage Monitoring and Defect Identification Based on Deep/Machine Learning)

► Show Figures

Figure 1

5 pages, 190 KiB

Open AccessEditorial

Machine Learning-Based Feature Extraction and Selection

by David Ruano-Ordás

Appl. Sci. 2024, 14(15), 6567; https://doi.org/10.3390/app14156567 (registering DOI) - 27 Jul 2024

Abstract

Over the last decade, technological advances have brought breakthroughs in the landscape of data management, transmission, processing, and storage [...] Full article

(This article belongs to the Special Issue Machine-Learning-Based Feature Extraction and Selection)

26 pages, 17816 KiB

Open AccessArticle

The Automated Generation of Medical Reports from Polydactyly X-ray Images Using CNNs and Transformers

by Pablo de Abreu Vieira, Mano Joseph Mathew, Pedro de Alcantara dos Santos Neto and Romuere Rodrigues Veloso e Silva

Appl. Sci. 2024, 14(15), 6566; https://doi.org/10.3390/app14156566 (registering DOI) - 27 Jul 2024

Abstract

Pododactyl radiography is a non-invasive procedure that enables the detection of foot pathologies, as it provides detailed images of structures such as the metatarsus and phalanges, among others. This examination holds potential for employment in CAD systems. Our proposed methodology employs generative artificial [...] Read more.

Pododactyl radiography is a non-invasive procedure that enables the detection of foot pathologies, as it provides detailed images of structures such as the metatarsus and phalanges, among others. This examination holds potential for employment in CAD systems. Our proposed methodology employs generative artificial intelligence to analyze pododactyl radiographs and generate automatic medical reports. We used a dataset comprising 16,710 exams, including images and medical reports on pododactylys. We implemented preprocessing of the images and text, as well as data augmentation techniques to improve the representativeness of the dataset. The proposed CAD system integrates pre-trained CNNs for feature extraction from the images and Transformers for report interpretation and generation. Our objective is to provide reports describing pododactyl pathologies, such as plantar fasciitis, bunions, heel spurs, flat feet, and lesions, among others, offering a second opinion to the specialist. The results are promising, with BLEU scores (1 to 4) of 0.612, 0.552, 0.507, and 0.470, respectively, a METEOR score of 0.471, and a ROUGE-L score of 0.633, demonstrating the model’s ability to generate reports with qualities close to those produced by specialists. We demonstrate that generative AI trained with pododactyl radiographs has the potential to assist in diagnoses from these examinations. Full article

(This article belongs to the Special Issue Recent Advances in and Applications of Medical Image Processing and Analysis)

► Show Figures

Figure 1

18 pages, 1302 KiB

Open AccessReview

Bridging the Gap: A Literature Review of Advancements in Obesity and Diabetes Mellitus Management

by Gheorghe Nicusor Pop, Felicia Manole, Florina Buleu, Alexandru Catalin Motofelea, Silviu Bircea, Daian Popa, Nadica Motofelea and Catalin Alexandru Pirvu

Appl. Sci. 2024, 14(15), 6565; https://doi.org/10.3390/app14156565 (registering DOI) - 27 Jul 2024

Abstract

This literature review explores advancements in obesity and diabetes mellitus diagnosis and treatment, highlighting recent innovations that promise more personalized and effective healthcare interventions. For obesity diagnosis, traditional methods like body mass index (BMI) calculations are now complemented by bioelectrical impedance analysis (BIA) [...] Read more.

This literature review explores advancements in obesity and diabetes mellitus diagnosis and treatment, highlighting recent innovations that promise more personalized and effective healthcare interventions. For obesity diagnosis, traditional methods like body mass index (BMI) calculations are now complemented by bioelectrical impedance analysis (BIA) and dual-energy X-ray absorptiometry (DXA) scans, with emerging biomarkers from “omics” technologies. Diabetes diagnosis has advanced with standard hemoglobin A1c (HbA1c) testing supplemented by novel measures such as advanced glycation end products (AGEs) and autoantibodies, alongside the use of artificial intelligence to enhance diagnostic accuracy. Treatment options for obesity are expanding beyond traditional methods. Minimally invasive bariatric surgeries, endoscopic procedures, fecal microbiota transplants (FMTs), and pharmaceuticals like GLP-1 receptor agonists (semaglutide, tirzepatide) show promising results. Cognitive behavioral therapy (CBT) and prescription digital therapeutics (PDTs) are also valuable tools for weight management. Diabetes treatment is also undergoing a transformation. Ultra-long-acting insulins and innovative oral insulin delivery methods are on the horizon. SGLT2 inhibitors and GLP-1 receptor agonists are proving to be effective medications for blood sugar control. Continuous glucose monitoring (CGM) systems and closed-loop insulin delivery are revolutionizing diabetes management, while stem cell therapy holds promise for the future. By integrating advanced diagnostic tools with personalized treatment plans, obesity and diabetes care are entering a new era. This personalized approach empowers patients and paves the way for improved health outcomes and a better quality of life. Full article

(This article belongs to the Special Issue Advancements in Obesity and Diabetes Management: From Diagnosis to Treatment)

► Show Figures

Figure 1

13 pages, 4133 KiB

Open AccessArticle

Gender Recognition Based on the Stacking of Different Acoustic Features

by Ergün Yücesoy

Appl. Sci. 2024, 14(15), 6564; https://doi.org/10.3390/app14156564 (registering DOI) - 27 Jul 2024

Abstract

A speech signal can provide various information about a speaker, such as their gender, age, accent, and emotional state. The gender of the speaker is the most salient piece of information contained in the speech signal and is directly or indirectly used in [...] Read more.

A speech signal can provide various information about a speaker, such as their gender, age, accent, and emotional state. The gender of the speaker is the most salient piece of information contained in the speech signal and is directly or indirectly used in many applications. In this study, a new approach is proposed for recognizing the gender of the speaker based on the use of hybrid features created by stacking different types of features. For this purpose, four different features, namely Mel frequency cepstral coefficients (MFCC), Mel scaled power spectrogram (Mel Spectrogram), Chroma, Spectral contrast (Contrast), and Tonal Centroid (Tonnetz), and twelve hybrid features created by stacking these features were used. These features were applied to four different classifiers, two of which were based on traditional machine learning (KNN and LDA) while two were based on the deep learning approach (CNN and MLP), and the performance of each was evaluated separately. In the experiments conducted on the Turkish subset of the Common Voice dataset, it was observed that hybrid features, created by stacking different acoustic features, led to improvements in gender recognition accuracy ranging from 0.3 to 1.73%. Full article

(This article belongs to the Special Issue Speech Recognition and Natural Language Processing)

► Show Figures

Figure 1

Journal Description

Applied Sciences

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