Search Results (838)

Artificial intelligence technologies, particularly machine learning and computer vision, are being increasingly utilized to preserve, restore, and create immersive virtual experiences with cultural artifacts and sites, thus aiding in conserving cultural heritage and making it accessible to a global audience. This paper examines the performance of Generative Adversarial Networks (GAN), especially Style-Based Generator Architecture (StyleGAN), as a deep learning approach for producing realistic images of Egyptian monuments. We used Sigmoid loss for Language–Image Pre-training (SigLIP) as a unique image–text alignment system to guide monument generation through semantic elements. We also studied truncation methods to regulate the generated image noise and identify the most effective parameter settings based on architectural representation versus diverse output creation. An improved discriminator design that combined noise addition with squeeze-and-excitation blocks and a modified MinibatchStdLayer produced 27.5% better Fréchet Inception Distance performance than the original discriminator models. Moreover, differential evolution for latent-space optimization reduced alignment mistakes during specific monument construction tasks by about 15%. We checked a wide range of truncation values from 0.1 to 1.0 and found that somewhere between 0.4 and 0.7 was the best range because it allowed for good accuracy while retaining many different architectural elements. Our findings indicate that specific model optimization strategies produce superior outcomes by creating better-quality and historically correct representations of diverse Egyptian monuments. Thus, the developed technology may be instrumental in generating educational and archaeological visualization assets while adding virtual tourism capabilities. Full article

Civil aviation safety remains a critical concern globally, with continuous efforts aimed at reducing accidents and fatalities. This paper focuses on the comprehensive evaluation of civil aviation safety in the Slovak Republic over the past several years, with the main objective of identifying prevailing trends and key risk factors. A comprehensive analysis of 155 accidents and incidents was conducted based on selected operational parameters. Logistic regression was applied to identify potential causal factors influencing various levels of injury severity in aviation accidents. Moreover, the prediction model can also be used to predict the probability of specific injury severity for accidents with given parameter values. The results indicate a clear declining trend in the annual number of aviation safety events; however, the fatality rate has stagnated or slightly increased in recent years. Human error, particularly mistakes and intentional violations of procedures, was identified as the dominant causal factor across all sectors of civil aviation, including flight operations, airport management, maintenance, and air navigation services. Despite technological advancements and regulatory improvements, human-related failures persist as a major safety challenge. The findings highlight the critical need for targeted strategies to mitigate human error and enhance overall aviation safety in the Slovak Republic. Full article

Current mainstream classification methods predominantly employ end-to-end multi-class frameworks. These approaches encounter inherent challenges including high-dimensional feature space complexity, decision boundary ambiguity that escalates with increasing class cardinality, sensitivity to label noise, and limited adaptability to dynamic model expansion. However, human beings may avoid these mistakes naturally. Research indicates that humans subconsciously employ a decision-making process favoring binary outcomes, particularly when responding to questions requiring nuanced differentiation. Intuitively, responding to binary inquiries such as “yes/no” often proves easier for humans than addressing queries of “what/which”. Inspired by the human decision-making hypothesis, we proposes a decision paradigm named the evolutionary binary decision framework (EBDF) centered around binary classification, evolving from traditional multi-classifiers in deep learning. To facilitate this evolution, we leverage the top-N outputs from the traditional multi-class classifier to dynamically steer subsequent binary classifiers, thereby constructing a cascaded decision-making framework that emulates the hierarchical reasoning of a binary decision tree. Theoretically, we demonstrate mathematical proof that by surpassing a certain threshold of the performance of binary classifiers, our framework may outperform traditional multi-classification framework. Furthermore, we conduct experiments utilizing several prominent deep learning models across various image classification datasets. The experimental results indicate significant potential for our strategy to surpass the ceiling in multi-classification performance. Full article

In this paper we explore ChatGPT’s ability to produce a summary, a precis, and/or an essay on the basis of excerpts from a novel—The Solid Mandala—by Nobel Prize Australian writer Patrick White. We use a number of prompts to test a number of functions related to narrative analysis from the point of view of the “sujet”, the “fable”, and the style. In the paper, we illustrate extensively a number of recurrent semantic mistakes that can badly harm the understanding of the contents of the novel. We made a list of 12 different types of semantic mistakes or parrotting we found GPT made, which can be regarded as typical for stochastic-based generation. We then tested Gemini for the same 12 mistakes and found a marked improvement in all critical key issues. The conclusion for ChatGPT is mostly negative. We formulate an underlying hypothesis for its worse performance, the influence of vocabulary size, which in Gemini is seven times higher than in GPT. Full article

Objectives: To investigate whether a Multimodal Rehabilitation Program (MRP) affects the change in visual–spatial abilities, especially attention, information-processing speed, visual–spatial learning, the severity of depression, and strategies for coping with pain in Complex Regional Pain Syndrome (CRPS) participants. Methods: The study was conducted between October 2021 and February 2023, with a 4-week rehabilitation program that included individual physiotherapy, manual and physical therapy, and psychological intervention such as psychoeducation, relaxation, and Graded Motor Imagery therapy. Twenty participants with CRPS and twenty healthy participants, forming a control group, were enlisted. The study was a 2-arm parallel: a CRPS group with MRP intervention and a healthy control group matched to the CRPS group according to demographic variables. Before and after, the MRP participants in the CRPS group were assessed for visual–spatial learning, attention abilities, severity of depression, and pain-coping strategy. The healthy control group underwent the same assessment without intervention before two measurements. The primary outcome measure was Reproduction on Rey–Osterrieth’s Complex Figure Test assessing visual–spatial learning. Results: In the post-test compared to the pre-test, the participants with CRPS obtained a significantly high score in visual–spatial learning (p < 0.01) and visual information-processing speed (p = 0.01). They made significantly fewer omission mistakes in visual working memory (p = 0.01). After the MRP compared to the pre-test, the CRPS participants indicated a decrease in the severity of depression (p = 0.04) and used a task-oriented strategy for coping with pain more often than before the rehabilitation program (p = 0.02). Conclusions: After a 4-week MRP, the following outcomes were obtained: an increase in visual–spatial learning, visual information-processing speed, a decrease in severity of depression, and a change in the pain-coping strategies—which became more adaptive. Full article

Quality assurance programs are critical to ensuring the consistency and reliability of point-of-care surveillance test results. In 2022, we launched Laos’ inaugural quality control (QC) and external quality assessment (EQA) program for national HIV recent infection surveillance. Our study aims to implement the first QC and EQA program for national HIV recent infection surveillance in Laos, utilizing non-infectious dried tube specimens (DTS) for quality control testing. This initiative seeks to monitor and assure the quality of HIV infection surveillance. We employed the Asante HIV-1 Rapid Test for Recent Infection (HIV-1 RTRI) point-of-care kit, using plasma specimens from the Thai Red Cross Society to create dried tube specimens (DTS). The DTS panels, including HIV-1 negative, HIV-1 recent, and HIV-1 long-term samples, met ISO 13528:2022 standards to ensure homogeneity and stability. These panels were transported from the Thai National Institute of Health (Thai NIH) to the Laos National Center for Laboratory and Epidemiology (NCLE) and subsequently shipped to 12 remote laboratories at ambient temperature. The laboratory results were electronically transmitted to Thai NIH 15 days after receiving the panel for performance analysis. The concordance results with the sample types were scored, and laboratories that achieved 100% concordance across all sample panels were considered to have satisfactorily met the established standards. Almost all laboratories demonstrated satisfactory results with 100% concordance across all sample panels during all three rounds of QC: 11 out of 12 (92%) in June, 10 out of 12 (83%) in July, and 11 out of 12 (91%) in August. The two rounds of EQA performed in June and August 2022 were satisfied by 8 out of 11 (72%) and 5 out of 10 (50%) laboratories, respectively. QC and EQA monitoring identified errors such as testing protocol mistakes and insufficient DTS panel dissolution, leading to improvements in HIV recency testing quality. Laboratories that reported errors were corrected and implemented further preventive actions. The QC and EQA program for HIV-1 RTRI identified errors in HIV recent infection testing. Implementing a specialized QC and EQA program for DTS marks a significant advancement in improving the accuracy and consistency of HIV recent infection surveillance. Continuous assessment is vital for addressing recurring issues. Full article

Environmentally modified housing [EMH; windowless, insulated sidewalls and ceiling, thermostatically controlled ventilation fans) versus conventional housing [CVH; cross-ventilated, insulated ceiling, ceiling fans) improved broiler performance in the summer. The objective of this investigation was to determine whether social behaviors differed between two population densities (0.06 m²/chick [HD] or 0.07 m²/chick [LD]) in these houses. We used a randomized block statistical design, involving houses, population densities, observation times, and bird age. Behaviors were observed weekly, during the morning and the afternoon. Individual observers focused on the group of broilers in one of three defined 26.76 m² areas in each of the four pens in each house. Aggressive encounters, tail and back pecking, feather eating, thermoregulatory, preening, and flock mobility were recorded. Feather pecking, eating and aggressive encounters were expressed at greater rates in HD birds in CVH. A salt-deficient diet caused increased feather pecking and aggressive encounters, which decreased after correction of the mistake. Increased heat indices (HIs), HD, and greater light intensity in CVH influenced behaviors and mortality more severely than in EMH. In CVH and EMH, burrowing/thermoregulatory/resting activity increased with increasing HIs. Afternoon preening was elevated significantly in EMH. It was concluded that broilers reared in EMH were more comfortable and experienced improved welfare compared to those reared in CVH. Full article

The wide availability of sensor data stored in multiple formats makes it difficult to reuse in other applications. We consider the problem of extracting sensor data from unstructured and semi-structured texts using Large Language Models. With careful prompt crafting, we have been able to establish a strict JSON structure which can be further processed with automated ease. We establish a workflow that enables the extraction of data using GPT-4, Llama 3, Mistral and Falcon models, and we show that while the closed-source GPT-4 model is generally leading in conversion efficiency, other open-source models can follow this if given appropriate data structures. We define new measures to simplify the comparison, and we present a multi-purpose workflow for sensor data extraction. We observe that some of the smaller models are incapable of correctly extracting data from freeform text but are skilled in processing tabular data. On the other hand, larger models are more robust and avoid conversion mistakes more easily. Full article

This paper explores the application of deep learning techniques for evaluating and classifying deadlift weightlifting techniques from video input. The increasing popularity of weightlifting, coupled with the injury risks associated with improper form, has heightened interest in this area of research. To address these concerns, we developed an application designed to classify three distinct styles of deadlifts: conventional, Romanian, and sumo. In addition to style classification, our application identifies common mistakes such as a rounded back, overextension at the top of the lift, and premature lifting of the hips in relation to the back. To build our model, we created a comprehensive custom dataset comprising lateral-view videos of lifters performing deadlifts, which we meticulously annotated to ensure accuracy. We adapted the MoveNet model to track keypoints on the lifter’s joints, which effectively represented their motion patterns. These keypoints not only served as visualization aids in the training of Convolutional Neural Networks (CNNs) but also acted as the primary features for Long Short-Term Memory (LSTM) models, both of which we employed to classify the various deadlift techniques. Our experimental results showed that both models achieved impressive F1-scores, reaching up to 0.99 for style and 1.00 for execution form classifications on the test dataset. Furthermore, we designed an application that integrates keypoint visualizations with motion pattern classifications. This tool provides users with valuable feedback on their performance and includes a replay feature for self-assessment, helping lifters refine their technique and reduce the risk of injury. Full article

Mining involves daily descents underground and enduring dangerous and difficult conditions. Hence, it is very important to use solutions that will reduce the risk in miners’ work and ensure the greater safety and comfort of work in accordance with the goals of sustainable development. One way is training using virtual reality. Virtual reality provides greater safety (safe training conditions, the possibility of making a mistake without health consequences, practicing emergency scenarios, etc.) and aligns with the Sustainable Development Goals—particularly SDG 3 (health), SDG 8 (decent work), SDG 9 (innovation), and SDG 12 (sustainable production). However, it is also a technology that has its weaknesses (occurrence of contraindications, side effects, etc.). Therefore, the use of VR-based training should be examined in terms of the well-being and health of training employees. Due to this, this article examines the occurrence of psychophysical complaints during VR training; the tolerance and adequacy of the duration of a 50 min training session in VR was assessed; and the average time needed to adapt to the virtual environment was determined. The VR training was developed as a result of a research project conducted by JSW Nowe Projekty S.A. (ul. Ignacego Paderewskiego 41, 40-282 Katowice, Poland), Główny Instytut Górnictwa—Państwowy Instytut Badawczy (plac Gwarków 1, 40-160 Katowice, Poland), JSW Szkolenie i Górnictwo Sp. z o.o. at Jastrzębska Spółka Węglowa Capital Group (ul. Górnicza 1, 44-335 Jastrzębie-Zdrój, Poland) on the development and implementation of innovative training using VR for miners. The solution was developed in the context of mining’s striving for sustainable development in the area of improving working conditions and human safety. The first method used in the study is a survey completed by participants of training courses using virtual reality. The second method is the analysis of trainer observation sheets, which contain observations from training courses. The results revealed that for over 70% of respondents, the need to carry out activities in VR was not associated with fatigue. No average score for psychophysical symptoms assessed by respondents on a scale of 1 to 6 (including disorientation, blurred vision, dizziness, confusion, etc.) exceeded 1.4. The vast majority (85.5%) did not take off the goggles before the end of the training—the training lasted 50 min. This research contributes to the discussion on sustainable industrial transformation by demonstrating that VR training not only improves worker safety and preparedness but also supports development goals through human-centered innovation in the mining sector. Full article

If artificial intelligence (AI) is to be applied in safety-critical domains, its performance needs to be evaluated reliably. The present study investigated how humans evaluate AI systems for person detection in automatic train operation. In three experiments, participants viewed image sequences of people moving in the vicinity of railway tracks. A simulated AI system highlighted all detected people—sometimes correctly and sometimes not. Participants had to provide a numerical rating of the AI’s performance and then verbally explain their rating. The experiments manipulated several factors that might influence human ratings: the types and plausibility of AI mistakes, the number of affected images, the number of people present in an image, the position of people relevant to the tracks, and the methods used to elicit human evaluations. While all these factors influenced human ratings, some effects were unexpected or deviated from normative standards. For instance, the factor with the strongest impact was people’s position relative to the tracks, although participants had explicitly been instructed that the AI could not process such information. Taken together, the results suggest that humans may sometimes evaluate more than the AI’s performance on the assigned task. Such mismatches between AI capabilities and human expectations should be taken into consideration when conducting safety audits of AI systems. Full article

Background: Stereolithography is a rapid prototyping and 3D printing technique that creates solid three-dimensional models. An accurate and functional 3D model using stereolithography is invaluable in scientific research, particularly in studies involving edentulous patients. Additive manufacture and CAD systems help achieve accurate measurements and procedures and be easily replicated by lowering human error mistakes. The main objective of this study was to develop an in vitro simulation model with a reduced alveolar ridge with the same characteristics as mandibular edentulous patients using stereolithography. Methods: A mandibular model with a resorbed mandibular crest was scanned, and the STL model was aligned to the XYZ reference system. A reduction in the alveolar ridge corresponding to the mandibular mucosa of an edentulous patient was achieved. A negative model also derived from the original model was made to ensure the space for oral simulation material. A dimensional stability test was performed to validate the model. Results: The maximal mean displacement of the model was 0.015 mm, and the minimal mean displacement was 0.004 mm. The oral mucosa had a displacement of approximately 1.6 mm. Conclusions: An in vitro 3D simulation model of a complete edentulous patient mucosa was achieved. Full article