Search Results (219)

Traditional approaches to safety event analysis in autonomous systems have relied on complex machine and deep learning models and extensive datasets for high accuracy and reliability. However, the emerge of multimodal large language models (MLLMs) offers a novel approach by integrating textual, visual, and audio modalities. Our framework leverages the logical and visual reasoning power of MLLMs, directing their output through object-level question–answer (QA) prompts to ensure accurate, reliable, and actionable insights for investigating safety-critical event detection and analysis. By incorporating models like Gemini-Pro-Vision 1.5, we aim to automate safety-critical event detection and analysis along with mitigating common issues such as hallucinations in MLLM outputs. The results demonstrate the framework’s potential in different in-context learning (ICT) settings such as zero-shot and few-shot learning methods. Furthermore, we investigate other settings such as self-ensemble learning and a varying number of frames. The results show that a few-shot learning model consistently outperformed other learning models, achieving the highest overall accuracy of about 79%. The comparative analysis with previous studies on visual reasoning revealed that previous models showed moderate performance in driving safety tasks, while our proposed model significantly outperformed them. To the best of our knowledge, our proposed MLLM model stands out as the first of its kind, capable of handling multiple tasks for each safety-critical event. It can identify risky scenarios, classify diverse scenes, determine car directions, categorize agents, and recommend the appropriate actions, setting a new standard in safety-critical event management. This study shows the significance of MLLMs in advancing the analysis of naturalistic driving videos to improve safety-critical event detection and understanding the interactions in complex environments. Full article

Since in the automotive field colors play an important role, the present study tried to answer the following questions: is the perceived quality (PQ) of the vehicle interior color different after visually exploring the car body color? If so, how? Here, exploiting immersive virtual reality simulations and eye-tracking technology, participants were asked to visually explore an unbranded car in different exterior/interior color combinations and rate its PQ. Fixation duration (time eyes are fixed on a target) was considered as an implicit measure of visual attention allocation while PQ evaluations were considered as explicit measures of individual preferences for car colors. As for eye-tracking data, the results showed that white and red car exteriors affected the attention to interiors with the fixation duration being longer for gray than black interiors. Moreover, the subjective evaluations of car PQ predicted eye-tracking patterns: as the negative evaluation increased, the fixation duration on car interiors also increased. Overall, these preliminary results suggested the need to further explore the relationship between PQ and attentional/motivational processing as well as the role of subjective aesthetic preferences for color combinations in the automotive field. Full article

This research examines student teachers’ perceptions of a game-based exam conducted in the Genial.ly app in the study course ”Legal Aspects of the Pedagogical Process”. This study aims to find out the pros and cons of game-based exams and understand which digital solutions can enable the development and analysis of digital game data. At the beginning of the course, students were introduced to the research and asked to provide feedback throughout the course on what they saw as the most important aspects of each class and insights on how they envisioned the game-based exam could proceed. The game-based exam was built using the digital platform Genial.ly after its update, which provided the possibility to include open-ended questions and collect data for analyses. It was designed with a narrative in which a new teacher comes to a school and is asked for help in different situations. After reading a description of each situation, the students answered questions about how they would resolve them based on Latvia’s regulations. After the exam, students wrote feedback indicating that the game-based exam helped them visualize the situations presented, resulting in lower stress levels compared to a traditional exam. This research was structured based on design-based principles and the data were analyzed from the perspective of how educators can use freely available solutions to develop game-based exams to test students’ knowledge gained during a course. The results show that Genial.ly can be used as an examination tool, as indicated by positive student teachers’ responses. However, this research has limitations as it was conducted with only one test group due to administrative reasons. Future research could address this by including multiple groups within the same course as well as testing game-based exams in other subject courses for comparison. Full article

This paper analyzes spatial gestures and cognition in a new, or so-called “emerging”, visual language, the Yucatec Maya Sign Language (YSML). This sign language was created by deaf and hearing signers in various Yucatec Maya villages on the Yucatec Peninsula (Mexico). Although the sign language is not a signed version of spoken Yucatec Maya, both languages evolve in a similar cultural setting. Studies have shown that cultures around the world seem to rely on one preferred spatial Frame of Reference (FoR), shaping in many ways how people orient themselves and think about the world around them. Prior research indicated that Yucatec Maya speakers rely on the use of the geocentric FoR. However, contrary to other cultures, it is mainly observable through the production of gestures and not speech only. In the case of space, gestures in spoken Yucatec Maya exhibit linguistic features, having the status of a lexicon. Our research question is the following: if the preferred spatial FoR among the Yucatec Mayas is based on co-expressivity and spatial linguistic content visually transmitted via multimodal interactions, will deaf signers of an emerging language created in the same cultural setting share the same cognitive preference? In order to answer this question, we conducted three experimental tasks in three different villages where YMSL is in use: a non-verbal rotation task, a Director-Matcher task and a localization task. Results indicate that YMSL signers share the same preference for the geocentric FoR. Full article

This research proposes an innovative, intelligent hand-assisted diagnostic system aiming to achieve a comprehensive assessment of hand function through information fusion technology. Based on the single-vision algorithm we designed, the system can perceive and analyze the morphology and motion posture of the patient’s hands in real time. This visual perception can provide an objective data foundation and capture the continuous changes in the patient’s hand movement, thereby providing more detailed information for the assessment and providing a scientific basis for subsequent treatment plans. By introducing medical knowledge graph technology, the system integrates and analyzes medical knowledge information and combines it with a voice question-answering system, allowing patients to communicate and obtain information effectively even with limited hand function. Voice question-answering, as a subjective and convenient interaction method, greatly improves the interactivity and communication efficiency between patients and the system. In conclusion, this system holds immense potential as a highly efficient and accurate hand-assisted assessment tool, delivering enhanced diagnostic services and rehabilitation support for patients. Full article

The significance of practical experience and visualization in the fluid power discipline, highly tied to students’ success, requires integrating immersive pedagogical tools for enhanced course delivery, offering real-life industry simulation. This study investigates the impact of using virtual reality (VR) technology as an instructional tool on the learning and engagement of 48 mechanical engineering technology (MET) students registered in the MET: 230 Fluid Power course at Purdue University. An interactive VR module on hydraulic grippers was developed utilizing the constructivist learning theory for MET: 230 labs, enabling MET students to explore light- and heavy-duty gripper designs and operation through assembly, disassembly, and testing in a virtual construction environment. A survey consisting of a Likert scale and short-answer questions was designed based on the study’s objective to evaluate the students’ engagement and perceived attitude toward the module. Statistical and natural language processing (NLP) analyses were conducted on the students’ responses. The statistical analysis results revealed that 97% of the students expressed increased excitement, over 90% reported higher engagement, and 87% found the VR lab realistic and practical. The NLP analysis highlighted positive themes such as “engagement”, “valuable experience”, “hands-on learning”, and “understanding”, with over 80% of students endorsing these sentiments. These findings will contribute to future efforts aimed at improving fluid power learning through immersive digital reality technologies, while also exploring alternative approaches for individuals encountering challenges with such technologies. Full article

The available research suggests that the rate of involvement of people with disabilities and their families in the life of the Church is significantly lower than that of people without disabilities. The engagement of people with disabilities is largely dependent on (a) the level of religiosity; (b) intrinsic motivation; (c) the level of trust in the institutions of the Church; and (d) broadly understood accessibility factors. Barriers experienced by people with disabilities are complex in nature, and make these people dependent on the help of others. Overcoming them requires significant investment, commitment, and change in the Church institution. These issues are relatively rarely addressed in the literature. The few, usually partial studies tend to concentrate on specific disabilities, discussed with no reference to the family context. However, it is usually the case that the religiosity and church activity of a person with a disability are firmly rooted in their family reality, shaped by the level of religiosity of their parents, and sometimes dependent on their presence and involvement. The aim of this article, which is both theoretical and empirical in nature, is to answer the question of which barriers form an obstacle to participation in religious life for people with disabilities and their families in Poland. This study uses the results of a nationwide qualitative research (focus group interview method) conducted among people with physical and intellectual disabilities, the hard-of-hearing and the deaf, the visually impaired, and their carers. Data analysis enabled the identification of the following barriers: infrastructural, personal and organizational (family-related and extra-familial). These research results can provide guidance in pastoral work with people with disabilities and their families, improving not only the quality of their religious experience, but also the number of the faithful in the Church community. Full article

“Are you LISTENING?” may be one of the most frequent questions preschoolers hear from their parents and teachers, but can children be taught to listen carefully—and thus better comprehend language—and if so, what changes occur in their brains? Twenty-seven four- and five-year-old children were taught a language simulation strategy to use while listening to stories: first, they practiced moving graphics on an iPad to correspond to the story actions, and then they practiced imagining the movements. Compared to a control condition, children in the intervention answered comprehension questions more accurately when imagining moving the graphics and on a measure of transfer using a new story without any instruction and with only immovable graphics. Importantly, for children in the intervention, the change in comprehension from the first to the sixth day was strongly correlated with changes in EEG mu and alpha desynchronization, suggesting changes in motor and visual processing following the intervention. Thus, the data are consistent with our hypothesis that a language simulation listening comprehension intervention can improve children’s listening comprehension by teaching children to align visual and motor processing with language comprehension. Full article

Where there are wind turbines, they become a dominant feature of the landscape. The landscape is one of the frequently identified types of impacts of these investments on the natural environment and people. Specially prepared methodologies are used to assess the impact of turbines on the landscape. No less important is the subjective feeling of residents because it can affect the social acceptance of these investments. The work answers questions about residents’ opinions on the impact of energy installations on the landscape. The results of the study, using chi-square, indicate that there is a relationship between the presence of windmills in the municipality and support for their development, as well as the evaluation of both the positive as well as the negative. Residents of a municipality where wind turbines have been around for more than a dozen years considered the introduction of a very strong visual stimulus as the most important negative impact on the landscape. Residents of a municipality without wind power considered interference with ecosystems as the most important negative impact. Full article

With the advent of next-generation sequencing (NGS), experimental techniques that capture the biological significance of DNA loci or RNA molecules have emerged as fundamental tools for studying the epigenome and transcriptional regulation on a genome-wide scale. The volume of the generated data and the underlying complexity regarding their analysis highlight the need for robust and easy-to-use computational analytic methods that can streamline the process and provide valuable biological insights. Our solution, aPEAch, is an automated pipeline that facilitates the end-to-end analysis of both DNA- and RNA-sequencing assays, including small RNA sequencing, from assessing the quality of the input sample files to answering meaningful biological questions by exploiting the rich information embedded in biological data. Our method is implemented in Python, based on a modular approach that enables users to choose the path and extent of the analysis and the representations of the results. The pipeline can process samples with single or multiple replicates in batches, allowing the ease of use and reproducibility of the analysis across all samples. aPEAch provides a variety of sample metrics such as quality control reports, fragment size distribution plots, and all intermediate output files, enabling the pipeline to be re-executed with different parameters or algorithms, along with the publication-ready visualization of the results. Furthermore, aPEAch seamlessly incorporates advanced unsupervised learning analyses by automating clustering optimization and visualization, thus providing invaluable insight into the underlying biological mechanisms. Full article

The university population has high visual demands. It is therefore important to assess the prevalence of symptoms in these subjects, which may affect their academic performance. In this cross-sectional study, a randomized sample of 252 subjects from a university answered the Convergence Insufficiency Symptom Survey (CISS) questionnaire. In addition, questions were asked about blurred vision during and after near tasks, the number of hours per day spent in near vision, and whether or not they wore glasses. Furthermore, 110 subjects underwent an eye exam, including a refraction and accommodation assessment. The mean age of the subjects was 28.79 ± 11.36 years, 62.3% reported wearing glasses, and on average 7.20 ± 2.92 hours/day was spent in near vision. The mean of the CISS score was 18.69 ± 9.96, and according to its criteria, 38% of the subjects were symptomatic. Some symptoms were significantly (p < 0.05) more frequent in subjects wearing glasses. Accommodative dysfunctions were present in 30.9% of the subjects, the most common being insufficiency of accommodation. We emphasise the importance of assessing symptomatology during the clinical examination in this group of subjects, as they spend many hours a day in near vision, as well as assessing accommodation, binocular vision, and the ergonomic work environment, which may be at the origin of the symptoms, in addition to the need to wear glasses. Full article

This study addresses the significant challenge posed by the use of Large Language Models (LLMs) such as ChatGPT on the integrity of online examinations, focusing on how these models can undermine academic honesty by demonstrating their latent and advanced reasoning capabilities. An iterative self-reflective strategy was developed for invoking critical thinking and higher-order reasoning in LLMs when responding to complex multimodal exam questions involving both visual and textual data. The proposed strategy was demonstrated and evaluated on real exam questions by subject experts and the performance of ChatGPT (GPT-4) with vision was estimated on an additional dataset of 600 text descriptions of multimodal exam questions. The results indicate that the proposed self-reflective strategy can invoke latent multi-hop reasoning capabilities within LLMs, effectively steering them towards correct answers by integrating critical thinking from each modality into the final response. Meanwhile, ChatGPT demonstrated considerable proficiency in being able to answer multimodal exam questions across 12 subjects. These findings challenge prior assertions about the limitations of LLMs in multimodal reasoning and emphasise the need for robust online exam security measures such as advanced proctoring systems and more sophisticated multimodal exam questions to mitigate potential academic misconduct enabled by AI technologies. Full article

The primary goal of this research is to identify which factors most significantly influence customer satisfaction in the last-mile delivery (LMD) process. The sample comprised 907 participants (63.4% female) with a mean age of 34.90. All participants completed three questionnaires regarding LMD, customer satisfaction, and trust in courier service. Furthermore, participants answered questions related to significant aspects of the delivery process: speed, price, and courier call before delivery. To determine which factors most significantly influence customer satisfaction in LMD, structural equation modeling (SEM) was applied. The tested SEM model showed a good fit. The results indicated that within the LMD dimension, visual appeal was a significant predictor in a negative direction, and all other LMD dimensions (except parcel tracking) were positive and significant predictors of customer satisfaction. Trust in courier service, delivery price, speed, and courier call before delivery were statistically significant predictors of customer satisfaction in last-mile delivery, all in a positive direction. Full article

Medical Visual Question Answering (Med-VQA) faces significant limitations in application development due to sparse and challenging data acquisition. Existing approaches focus on multi-modal learning to equip models with medical image inference and natural language understanding, but this worsens data scarcity in Med-VQA, hindering clinical application and advancement. This paper proposes the ITLTA framework for Med-VQA, designed based on field requirements. ITLTA combines multi-label learning of medical images with the language understanding and reasoning capabilities of large language models (LLMs) to achieve zero-shot learning, meeting natural language module needs without end-to-end training. This approach reduces deployment costs and training data requirements, allowing LLMs to function as flexible, plug-and-play modules. To enhance multi-label classification accuracy, the framework uses external medical image data for pretraining, integrated with a joint feature and label attention mechanism. This configuration ensures robust performance and applicability, even with limited data. Additionally, the framework clarifies the decision-making process for visual labels and question prompts, enhancing the interpretability of Med-VQA. Validated on the VQA-Med 2019 dataset, our method demonstrates superior effectiveness compared to existing methods, confirming its outstanding performance for enhanced clinical applications. Full article

Liturgical space represents the embodiment of Christian theology in church buildings, encompassing both physical and metaphysical aspects. This space carries holiness and sacredness through a set of architectural elements that create sacred and profane zones within the church architecture. For centuries, design elements have shaped the form of Eastern churches in Iraq. This research aimed to answer the following question: what does a participant see at first glance in the liturgical space of a church building? This paper revisits the impact of design elements on the liturgical space of Eastern churches. The research methodology involved analyzing qualitative data using visual attention software (VAS) 3M, version 2024, to examine eye-tracking data and identify what visitors first noticed when entering these church interiors in Mosul, Iraq. The results highlight the variations and dominance of specific design elements in their impact on Eastern churches. The conclusions emphasize the importance of scientifically based restoration for the perception of design elements in these churches. Full article