In recent years, there have been frequent incidents of foreign objects intruding into railway and Airport runways. These objects can include pedestrians, vehicles, animals, and debris. This paper introduces an improved YOLOv5 architecture incorporating FasterNet and attention mechanisms to enhance the detection of foreign objects on railways and Airport runways. This study proposes a new dataset, the aero and rail foreign object detection (AARFOD), which combines two public datasets for detecting foreign objects in aviation and railway systems. The dataset aims to improve the recognition capabilities of foreign object targets. Experimental results on this large dataset have demonstrated significant performance improvements of the proposed model over the baseline YOLOv5 model, reducing computational requirements. Improved YOLO model shows a significant improvement in precision by $1.2 \%$, recall rate by $1.0 \%$, and mAP@. 5 by $0.6 \%$, while mAP@. $5-.95$ remained unchanged. The parameters were reduced by approximately ${2 5. 1 2 \%}$, and GFLOPs were reduced by about $10.63 \%$. In the ablation experiment, it is found that the FasterNet module can significantly reduce the number of parameters of the model, and the reference of the attention mechanism can slow down the performance loss caused by lightweight.
Delicate cloth simulations have long been desired in computer graphics. Various methods were proposed to improve engaged force interactions, collision handling, and numerical integrations. Deep learning has the potential to achieve fast and real-time simulation, but common neural network (NN) structures often demand many parameters to capture cloth dynamics. This paper proposes a physics-embedded learning framework that directly encodes physical features of cloth simulation. The convolutional NN is used to represent spatial correlations of the mass-spring system, after which three branches are designed to learn linear, nonlinear, and time derivate features of cloth physics. The framework can also integrate with other external forces and collision handling through either traditional simulators or sub NNs. The model is tested across different cloth animation cases, without training with new data. Agreement with baselines and predictive realism successfully validate its generalization ability. Inference efficiency of the proposed model also defeats traditional physics simulation. This framework is also designed to easily integrate with other visual refinement techniques like wrinkle carving, which leaves significant chances to incorporate prevailing machine learning techniques in 3D cloth amination.
The developments of artificial intelligence (AI) are growing along with its applications. This growth is so quick that it often surprises even researchers who had hypothesized different times. Within the field of criminal profiling, this is interesting because it can help to recognize errors and biases that are typical of humans [1]. Even though training AI to recognize emotions based on biometric parameters is becoming easier, the subsequent analyses are problematic. In fact, it is difficult to interpret biometric data, which are also influenced by cultural and social factors. In terrorism analysis, for instance, the behaviors that are analyzed are different among the different groups or tribes. Therefore, the influence of social factors goes beyond the analysis of the complex neural responses [2–4]. Another element that plays a role is in the interpretation of emotions for the judicial system, which is represented by ethical and moral factors [5]. Artificial intelligence cannot be used for reconstructing the origin of a crime [6] and only an expert’s opinion can be considered reliable [7]. Only an analysis based on the individual and aspects, and only the knowledge of the psychopathology, together with the scientific analysis of the nonverbal language, can help reconstruct the origin and the dynamics of the crime [10–12]. In conclusion, even though AI offers an important support since it can speed up some processes of the analysis, it currently cannot replace humans when it comes to profiling [13, 14]. In light of the chosen method, the analyses are ongoing, and the initial results indicate a trend toward greater reliability for profiling conducted by a human compared to that performed by AI. This is not due to the AI’s capacity for emotional recognition but rather to the methodology employed by the AI. Humans respond to any sensory stimulation with an emotion, making any inference, reasoning, or behavioral choice closely dependent on the emotion experienced. In contrast, AI recognizes emotions through a process of analysis comparable to purely cognitive processes. Consequently, the capacity for emotional recognition through empathy is lacking. To guarantee the best possible analysis and limit the possibility of moral and ethical issues, it is extremely important for a human to oversee this process. AI can be used to recognize emotions based on biometric alterations, but it should not go further than that. Relying solely on its conclusions would be sterile and incomplete, and from a legal standpoint, could impact the admissibility of the analysis in court.
In this paper, we have considered the coverage problem in wireless sensor network (WSN) on a convex subset of $R^{2}$. Sensors are dropped from the air randomly on some pre-fixed points, which is known as vertices, of region of interest (ROI). We use optimal partition of the ROI, which is actually partition in several regular hexagons. Since sensors are distributed randomly, a sensor may not be placed on the target vertex. For this reason, ROI will not be completely covered by a set of sensors. In practice, few more sensors are deployed on few (randomly chosen) vertices or used actuator (it can carry sensors to the proper vertex) to reduce the uncovered region or area. In one of our previous works, we have developed a strategy as follows: reduce the distance among two adjacent vertices and deployed one sensor on a vertex so that total number of sensors will be same as in existing old method (drop two sensors on some vertices and one sensor on the rest). We have compared the proportion of uncovered region using the commonly used old strategy with our previous one. We have simulated for several values of percentage of extra sensors and observed that our previous strategy is better for low standard deviation (s.d.), but not better for higher s.d. in both two and three dimensions. Inspiring from the above fact, in this paper, we combined above two strategies to find a general one, for deploying sensors in two dimensions. The excess sensors are divides in two parts. One part is used for decrease the side of the regular hexagon and other part is used for using one more sensor on some selected points. We simulate uncovered area and results indicate the optimal choice of these two parts, which change with the standard deviation of randomness.
This study focuses on the vulnerabilities and attack vectors connected with ransomware in Elastic Sky X integrated (ESXi) settings. We offer a novel technique to address these concerns by mimicking an ESXi environment, focusing on honeypot deployment and monitoring behaviors. Our strategy is creating a controlled emulation of ESXi in which we place honeypots to lure and capture ransomware activity. Furthermore, we use sophisticated monitoring methods to watch and evaluate ransomware behaviors in this simulated environment. Our approach’s effectiveness is tested using the simulated ESXi environment’s detection and response capabilities. The findings show that using honeypots in conjunction with careful behavioral monitoring can considerably improve the identification and mitigation of ransomware threats in virtualized environments.
In academic writing, the accuracy of citation formatting in scientific publications is essential for maintaining the integrity and consistency of scientific communication. However, manually formatting citations according to different styles, such as the IEEE, APA, or MLA, can be time-consuming and error-prone. This paper presents an innovative approach to automate citation formatting in scientific publications using ChatGPT. It proposes an algorithm that incorporates a sequence of instructions and guidance, combined with the capabilities of ChatGPT, and greatly simplifies the process of formatting citations according to different styles. The proposed approach involves training ChatGPT with a dataset containing citation guides and examples from different formatting styles to improve its ability to generate correctly formatted citations. This work presents a comparative characterization of the existing automated citation formatting systems and the proposed algorithm with ChatGPT. Their functionalities are analyzed and their advantages and disadvantages are highlighted. In addition, a SWOT analysis of the systems is performed, which examines their strengths, weaknesses, opportunities, and threats. The analysis highlights the effectiveness and advantages of the proposed solution with ChatGPT. The results show that automation using ChatGPT not only facilitates accurate citation formatting but also offers a practical tool for improving the quality and relevance of scientific publications. ChatGPT can significantly reduce formatting errors and improve the efficiency of academic writing, offering a scalable solution for researchers and institutions.
Reliable direction of arrival (DOA) estimation is crucial for the performance of wireless communication systems. In this paper, we introduce a refined DOA estimation method that combines eigenvalue reconstruction of the noise subspace and Toeplitz preprocessing with the multiple signal classification (MUSIC) algorithm. The proposed technique enhances the consistency of the noise subspace and improves the algorithm’s resolution. Extensive simulations demonstrate that the method outperforms both the standard MUSIC and the MUSIC with Eigenvalue Reconstruction (MUSIC_ER) techniques. Notably, our approach shows enhanced performance in terms of the root mean square error (RMSE) across snapshot ranges from 1 to 10. These enhancements make the proposed method (MUSIC_TR) a practical and effective option, especially in low-snapshot scenarios, providing an alternative solution for DOA estimation.
In recent years, telehealth and telerehabilitation have been on the rise due to lockdown and quarantine placing restrictions on in-person healthcare during the 2020 COVID-19 Pandemic. However, there are some limitations to the service that physicians are able to do remotely. This paper proposes a contemporary way of collecting more data from the remote patient to help during their telerehabilitation appointments. Surface electrodes are placed on the participants’ forearm flexor muscles, and samples were collected for each exercise. The data were then analyzed, and the features-RMS, IEMG, and VARwere used. The exertion of the strength of the muscle during an exercise could be seen from plotting the RMS of the exercise to the IEMG time domain graph, and the classification of the exercise could be interpreted from the IEMG data from one exercise, that has been normalized, which was plotted as a box plot diagram to be compared with the other exercises. The findings from this paper could be used in helping to build a model to ensure that the exercises are done correctly and the muscles are not being strained too hard during an exercise by the physician during a telerehabilitation session.
Technology enables manufacturing small-and medium-sized enterprises (SMEs) to improve operational efficiency through the use of business management software, digital inventory systems, and automation of production processes. Technology opens up wider market access, allowing SMEs to reach potential consumers through digital platforms and e-commerce to international markets. This study aims to measure the readiness of manufacturing SMEs in Indonesia in implementing digital technology. This study uses Maturity Assessment with a questionnaire instrument. The findings show the average manufacturing SMEs in Indonesia is at the “Learning” readiness level with a percentage of $87 \%$ in which the textile and food sectors are recorded as the most ready to adopt digital technology. The findings also show that a higher category of technology readiness for SMEs is associated with the larger proportion of SMEs that participated in the government program.
The quality of commutator surfaces in DC motors significantly affects the performance and longevity of the motors. Traditional methods of inspecting commutator surface defects, such as roundness and roughness meters, have limitations in detecting subtle and complex surface irregularities. This study proposes an image analysis technique combined with convolutional neural networks to enhance the detection of commutator surface defects. Our method improves the identification and classification of defects, correlating these findings with the assembly quality of DC motors. Although the experimental results are premilitary, it validates the effectiveness of the proposed approach, demonstrating improvements in defect detection accuracy. Future work will focus on expanding the image dataset and refining the CNN model to enhance its accuracy and real-time application capabilities.
With the rapid growth of online transactions and interactions, the threat landscape of scams and fraud has evolved, necessitating sophisticated detection mechanisms. This paper provides an extensive review of the latest advances in detecting online scams and fraud, covering technological solutions, machine learning techniques, and emerging trends in the field. Key methods discussed include advanced machine learning algorithms for anomaly detection, user behavior analytics, and the integration of threat intelligence. Additionally, this study highlights the role of public awareness and education in preventing scams, as well as the importance of international collaboration in law enforcement. By examining current trends and emerging technologies, this study provides strategies for organizations and individuals to enhance their digital security posture, effectively mitigating the risks associated with online scams and frauds.
Optical advances in skincare technology represent a revolutionary approach to addressing various dermatological concerns and enhancing overall skin health. This study provides an in-depth exploration of the principles, applications, and benefits of optical technologies in skincare. From noninvasive diagnostics to targeted treatments and cosmetic formulations, optical innovations are transforming the landscape of skincare, offering new possibilities for personalized and effective solutions. Optical advances in skincare technology have the potential to transform dermatological practice and improve skin health outcomes for individuals worldwide.
