3D human pose estimation-based action recognition method for complex industrial scenarios

A low-latency deep learning approach for human action recognition in medical internet of things applications

A number of studies have shown the importance of combining observation with practice when learning an action. Although the Point-Light Display (PLD) technique is frequently used in laboratory contexts to assess the mechanisms underlying observational learning, it has only been infrequently used in ecological contexts. The present study explored whether the components of complex weightlifting movements could be learned by means of an observational learning protocol. To that end, we compared three observation conditions: action observation (Video group), Point-Light Display observation (PLD group), and no human action observation (Control group). Twenty-six athletes participated in a weekly one-hour session for 5weeks to learn a complex weightlifting movement. During this period, the participants alternated between phases of movement observation, which varied depending on the condition to which they were assigned, and phases of movement execution. There were 12 sets consisting of 2min of observation and 6 physical repetitions per session. Joint angles at key moments, bar trajectories, and intersegmental coordination were assessed both before (Pre-test) and after (Retention tests) the learning period. The results indicate that both observation conditions had a larger effect on learning than the control condition. Furthermore, the PLD condition was more effective than the video condition for complex intersegmental coordination. This experiment therefore suggests that PLD observation combined with physical practice can be beneficial for the acquisition of complex movements.

The páramo is a high-altitude ecosystem characterized by its herbaceous vegetation and distribution in tropical and subtropical regions. This ecosystem is highly sensitive to environmental disturbances, making it a priority area for conservation and research due to its biodiversity and strategic ecosystem functions. Consequently, it is essential to conduct ecological and conservation studies of páramo areas within interdisciplinary frameworks that address the various environmental, geopolitical, economic, and sociocultural challenges. The objectives of this study were: (1) to identify the richness of páramos and the evolution of knowledge in research during the period from 2014 to 2023; and (2) to determine the scientific output, keyword cooccurrence in articles, and the most influential researchers in the field during the period from 2014 to 2023. The methodology employed was descriptive bibliometric analysis, involving a comprehensive search for scientific articles in the Scopus database. For parameter visualization, VOSviewer and the Bibliometrix package in R Studio were used to apply Lotka’s Law. The results show that Colombia is the leading country in scientific production in this field, playing a central role in advancing knowledge about the páramo. Furthermore, the findings indicate that the impact of climate change and intensive human activities (such as agriculture, grazing, pine plantations, and tourism) have increased the risk of páramo degradation, altering hydrological cycles and reducing its regulatory capacity. This bibliometric study provides a robust foundation for the planning of public policies aimed at conservation, sustainable water resource management, and biodiversity protection in páramo ecosystems. Therefore, it is crucial to promote research that considers the páramo as a socio-ecological system, analyzing the interactions between human actors and the natural environment, which will enable the design of more equitable and effective management policies.

Purpose This article explores how architectural laboratories can function as sites of hybrid practice, reuniting architects with material agency. It critiques the reductionist translation of matter into abstract codes and standardized resources, proposing laboratories as spaces where epistemic extraction (knowledge production) and ontological experimentation (redefining material-human relations) converge to challenge extraction-based paradigms in architecture. Design/methodology/approach Through critical historical analysis and contemporary case studies, from Frederick Kiesler's mid-20th-century Design Correlation Laboratory to François Roche's New Territories practice, this article traces how laboratories have alternately perpetuated and disrupted architecture's abstracted relationship with matter. The research synthesizes new materialist theory (Jane Bennett's vibrant matter), planetary design pedagogy (Diana Agrest's nature-culture entanglement) and laboratory studies (Bruno Latour) in order to theorize hybrid laboratory practices that resist reductionist material epistemologies. Findings This article identifies a crucial distinction between extractive laboratories (which isolate material properties for universal application) and laboratories of entanglement (which activate situated, collaborative material relations). It demonstrates that architectural laboratories can operate as hybrid practices by acknowledging distributed agency across human and nonhuman actors, embracing speculative methodologies that value material unpredictability, developing representational innovations that make material agencies visible and generating project-specific codes more than applying universal standards. Historical and contemporary examples reveal how laboratories enable architects to design with. Originality/value Design and laboratory studies into a coherent framework for architectural practice; historically, it offers an alternative genealogy of architectural laboratories focused on material collaboration; practically, it articulates “laboratories of entanglement” as a specific model for hybrid practice that transgresses theory/practice, human/nonhuman and extraction/collaboration binaries. The concept of laboratories as sites of both epistemic extraction and ontological experimentation provides a novel lens for understanding how architectural knowledge production can resist reductionism while remaining materially grounded.

ABSTRACT This theoretical paper departs from the premise that local forms of adventure are often at risk of being erased or appropriated by colonial-capitalist systems such as tourism or the adventure sport industry. This phenomenon usually sidelines the needs of local communities and jeopardises the health of local ecosystems in the process. In suggesting that part of the problem lies in how the connection between adventure, place and culture is often theorised, the author goes on to introduce the term ‘Adventure Monuments’ – a concept that frames places of adventure as materially vibrant and challenges capitalist narratives that primarily view adventure as a human spectacle or commodity. Framing places as adventure monuments centres their material significance as relational and political spaces where adventure has been slowly co-authored by human and nonhuman agents. This approach draws from posthuman discourse to frame adventure, place and culture as mutually constitutive agencies that jointly evolve and become within a complex weave of relations involving multiple human and non-human actors. After its key theoretical premises are discussed, this perspective is applied to a case (The Trampolino of Rhodes) to demonstrate its epistemological value and how it can contribute to the wider effort to decolonise adventure at both local and global levels.

Earlier disaster research has often assumed that the primary agent of choice is the autonomous human individual, overlooking how such decisions are shaped through relationships, notions, infrastructure, technologies, institutions, residential choice, and non-human environments. Following two major flood disasters in 2019 and 2021, local volunteer groups, municipal officials, and residents in Takeo City, Saga Prefecture engaged in various collaborative efforts to rebuild livelihoods, navigate administrative systems, and prepare for future events. This study examines how Japan’s emerging basin-based flood management policy, Ryuiki-chisui (River Basin Disaster Resilience and Sustainability by All), is implemented and reinterpreted in everyday practice through an ethnographic case study of Takeo City. Drawing on “logic of care,” “saikan (in-between disaster),” and “fluid,” the study analyzes how flood governance unfolds not as a sequence of rational choices or top-down directives, but as ongoing socio-technical tinkering shaped by relational labor, uncertainty, and situated forms of expertise. The findings show that Ryuiki-chisui operates as a dynamic set of practices that bridge the gaps between hydrological models, administrative frameworks, and residents’ embodied knowledge of the Rokkaku River basin. Volunteer intermediaries are critical in translating institutional categories, coordinating support for disaster certification and emergency repairs, and addressing the grey zones that fall between formal systems. Through these practices, flood governance has become a care-intensive process involving continuous adjustments across human and non-human actors such as pumps, gates, tides, homes, legal documents, and community networks. By highlighting the relational and ethically charged dimensions of life amid recurring disasters, this study advances international discourse on inclusive and transdisciplinary water governance. It demonstrates that effective basin-based flood management relies not only on technical measures, but also on cultivating forms of collaboration and care that sustain communities within continually changing environments. Practically, the findings suggest that Ryuiki-chisui will be more effective when intermediary work for translation, coordination, and grey-zone problem solving is recognized and resourced as part of basin governance, alongside conventional hard and soft measures. Conceptually, the study provides a practice-oriented specification of socio-hydrological coupling that can inform future interdisciplinary research integrating socio-hydrological modeling with ethnographic and participatory approaches.

The recognition of unimodal behavioral postures is easily influenced by the environment, such as lighting, occlusion, etc. Therefore, most of the existing action recognition focuses on the multimodal field. Therefore, this article selects two typical methods at present for comparison, analysis and summary. Specifically, two methods, namely the dual-stream cross-modal fusion transformer and the model-based multimodal neural network, were selected for research. Among them, the core contribution of the dual-stream cross-modal fusion transformer is to achieve the fusion of RGB and depth information through modal enhancement and modal interaction. Model-based multimodal neural networks integrate OpenPose skeleton data with RGB frames and employ graph convolution and spatio-temporal ROI fusion. These two methods present two completely different approaches to action recognition in multimodal situations. In addition, this article also presents new research directions for the futureSuch as lightweighting, etc. And a specific application analysis was provided.

Event Stream based Human Action Recognition: A High-Definition Benchmark Dataset and Algorithms

Linguistic agency assignment is an unavoidable yet understudied feature in climate messaging. We propose temporal (moving-time/moving-ego), causal (natural/human), and action (non-agential/human) agency as distinct features targeting three core cognitive demands in climate communication: risk perception, responsibility, and efficacy beliefs. Through a 2 (temporal agency: moving-time vs. moving-ego) × 2 (causal agency: natural vs. human) × 2 (action agency: non-agential vs. human) between-subjects experiment with 1,917 participants, the results showed that moving-time temporal agency amplifies risk perceptions, human causal agency strengthens responsibility, and human action agency increases efficacy. However, inconsistent agency assignments reduce intention and trigger defensive reactions.

Abstract The emergence of generative artificial intelligence (GenAI) is reshaping the research landscape and carries significant implications for Digital Humanities (DH), a field long intertwined with computational methods and technologies. This study examines how DH scholars are adopting and critically evaluating GenAI in their research. Drawing on an international survey of 76 respondents and 15 in‐depth interviews, we investigate scholars' motivations for using GenAI tools, the specific practices through which they integrate these tools into their research, and their perceptions of the benefits, risks, and challenges associated with GenAI. Our findings reveal divergent opinions and imaginaries within the DH community: while many scholars view GenAI as a means to enhance efficiency and support reskilling, others express concern about its impact on scholarly identity, intellectual labor, and disciplinary values. Situated within the history of DH and analyzed through the lens of actor‐network theory, the results suggest that GenAI is being incrementally enrolled into DH research networks, reshaping relationships among human and nonhuman actors in ways that remain contested and actively negotiated. As one of the first empirical studies on this topic, this work provides an initial foundation for understanding GenAI's evolving role in DH scholarship and points toward avenues for future research.

This research systematically investigates the roles of effort and ability exhibited by robots during service encounters. Although both are essential to successful service delivery, effort has traditionally been regarded as a uniquely human attribute, and robot ability has often been treated as a single construct. To address this limitation, the current research conceptualizes ability as a multidimensional construct comprising absolute and relative abilities, which may operate through distinct evaluative mechanisms. Across two studies, the findings show that human-like movement serves as a behavioral cue that simultaneously evokes perceptions of effort as well as absolute and relative abilities. Importantly, even for non-humanoid robots lacking human appearance, movement patterns resembling human actions during task execution can elicit inferences about effort and ability. These perceptions increase perceived product quality, such as the expected taste of coffee prepared by a robot server, with two ability types playing distinct roles depending on prior experience.

The re-evaluation of education has been one of the vital concerns of the previous century, owing to the events therein that defined subsequent history. Various educational systems are said to have influenced the prevalence of colonial mindset alongside structures that hinder the autonomy of students vis-à-vis their freedom in learning. It is in this context that Paulo Freire develops his acclaimed pedagogy which promotes a humanising education that affirms persons as human beings. Over the manifold forms of oppression that threaten personal freedom, Freire’s pedagogy forwards dialogic education through which students consciously take part in personal and societal transformation. Although Freirean thought is associated with Marxism, the Thomistic tradition affirms its core which seeks ‘the vocation of becoming more fully human.’ Jacques Maritain clarifies that education’s primary aim is to ‘form a man,’ integrating into a person the cultural, social, and intellectual aspects of human life. Parallel to the Freirean emphasis on the complementarity of reflection and action in dialogic education is the Thomistic valuation for contemplation brought to completion by human action.

Background: Exploring intentions gives us deep insights into moral philosophy and the foundations of ethical decision-making in healthcare. The concept of intention (Niyyah) occupies a central position in Islamic moral philosophy and jurisprudence, serving as the primary criterion for evaluating the ethical status of human actions. Despite its centrality to Islamic bioethics, the theoretical framework of Niyyah and its practical applications in contemporary clinical dilemmas remain underexplored in the medical literature.Methods: This article represents a conceptual and ethical analysis of intention in medical practice, particularly within Arab and Islamic cultural contexts. I reviewed philosophical literature on intention, examining both Western ethical theories (particularly Kantian ethics) and Islamic moral philosophy. Islamic sources consulted include the Qur’an, authenticated Hadith collections, and classical and contemporary fiqh literature. I examined contemporary Islamic bioethics scholarship and analyzed specific medical ethics cases where intention plays a crucial role: assisted dying, medical experimentation, and futile care. I also examined historical case studies (the opioid crisis and retrolental fibroplasia) to show how well-intentioned medical interventions can lead to harmful outcomes.Results: In this article, I make a thorough examination of the concept of intention and related concepts. I discuss several medical dilemmas where intention plays a crucial role: the intention to treat principle, assisted dying, medical experimentation, and the conduct of futile care. Additionally, I examine how good intentions have sometimes led to harmful outcomes, using the opioid crisis and retrolental fibroplasia as cautionary examples. The article also addresses cultural considerations in Arab healthcare contexts, including traditional healing practices, pain management approaches influenced by stoicism and religious beliefs, and the importance of family involvement in medical decision-making.Conclusion: The Islamic concept of Niyyah reminds us that the moral worth of our actions depends not only on their outcomes but also on the purity and sincerity of our intentions. The Maqasid al-Shariah framework provides a valuable tool for evaluating whether our actions truly serve the higher purposes of preserving life, preventing harm, and maintaining human dignity. Understanding the role of intention in ethical decision-making is essential for healthcare providers serving Arab and Muslim populations, as Islamic values significantly influence patient preferences, family dynamics, and clinical decision-making processes.

ABSTRACT This paper describes the theoretical principles underpinning the pedagogical affordances of experiential learning through digital games, such as Minecraft, in order to engage primary school children in climate change challenges. It discusses the instructional value of an online interactive game for exploring possible climate change scenarios based on real-world data. The paper describes the instructional design of a series of school-based workshops, targeting students’ understanding of complex environmental systems and the interconnectedness of human actions and climate outcomes. The proposed design defines the expected learning outcomes of a game-based learning approach to sustainability, including problem-solving, collaboration, and climate literacy, as well as climate action skills. Participatory approaches through iterative cycles of collaboration and feedback were used in the co-construction of the instructional design. Stakeholders included children, teachers and teacher educators who collaborated with the research team of interdisciplinary voices relevant to climate change education, such as educationalists, environmental scientists, computer scientists, architects and urban planners. The instructional design aims to highlight novel ways of utilising digital technologies to foster local, actionable, tangible and engaging actions among children. In doing so, it provides a blueprint for educators interested in engaged digital pedagogies for teaching and learning about climate change.

Climate change particularly changes in temperature and rainfall patterns have serious long-term influences, mainly caused by human actions. It can leads to severe events like heat-waves, floods, increasing seas, and more extreme weather, which in turn affects across various aspects of socioeconomic factors, ecosystems, hydrology, health, and global warming. Temperature and rainfall are essential climatological variables that are being comprehensively studied across the Karachi city to understand and manage their dynamic nature. The objective of this study is to analyze the trend, magnitude, and change points in climatic variables to investigate their variability in the urban city Karachi. Rainfall and temperature datasets for the Karachi station from 1980 to 2020 was used. Monthly and annual precipitation as well as temperature (maximum, and minimum) was analyzed for possible trends using nonparametric Mann-Kendall (MK) test, while the Sen’s slope (SS) estimator was used for magnitude of a trend. The Pettitt’s test was applied to detect the abrupt change point in climatic variables. The results revealed increasing trend in annual and monthly maximum temperatures while for minimum temperature all the months illustrated significant increasing trend in the study period. The SS estimator revealed that the annual temperature increased with rate of 0.025 °C per year for the certain time interval. However, in contrast to variation in the temperature trend, the mean annual rainfall of the study period was observed 180.4 mm with no significant increasing trend while 366.8 mm rainfall was observed in the month of the August 2020. The only month of the February has statistically significant decreasing trend with rate of change of 0.097mm/year and no significant trends was present for all other months. This study provides evidence that can aid policy adaptation, and addressing climate change by responding extreme events like floods and heat-waves.

Abstract The Anthropocene has become a significant framework for understanding the scale and impacts of human activity on Earth (Crutzen, 2002; Steffen et al., 2007). Whether or not it is officially recognised as a geological epoch, the term denotes a planetary state in which human actions destabilise climate systems, cause biodiversity loss, and raise urgent questions of justice, responsibility, and care (Haraway, 2016). This article, adapted from a keynote delivered at the Scottish Educational Research Association ( sera ) conference in Dundee in December 2024, examines how education can respond to socio-ecological change in this context. Focusing on environmental pedagogy (or ecopedagogy), it argues that effective educational responses to the Anthropocene must move beyond the transmission of information to cultivate systemic literacy, ecological responsibility, and justice-oriented practice. These orientations are framed as conditions of learning in the Anthropocene, shaping how learners and communities engage with uncertainty, interdependence, and uneven geographies of power. Drawing on the Natural Environment Research Council ( nerc )-funded gallant (Glasgow as a Living Laboratory for Accelerating Novel Transformation) project, the article explores how place-based, participatory approaches can support learning that connects knowledge-making with action and advocacy. The paper argues that while education cannot resolve the challenges of the Anthropocene, it can equip learners and communities to navigate systemic crises and to build capacities for resilience, care, and solidarity in the face of planetary change.

Human Activity Recognition (HAR) is vital in healthcare, security, and human-computer interaction. Traditional unimodal or bimodal approaches often fail to capture the full complexity of human actions. To overcome this, TriNet-HAR, a novel trimodal deep learning framework is proposed that integrates depth maps, 3D skeletal joints, and inertial signals via a unified, explainable architecture. Each modality is processed using a dedicated module: Convolutional Neural Network (CNN) for depth, joint-wise attention layers mimicking Graph-Convolution-Network (GCN) for skeletons, and positional encoding-enhanced Long Short-Term Memory (LSTM) for Inertial Measurement Unit (IMU) data. A multi-head self-attention and cross-attention fusion strategy refines inter-modality dependencies. SHapley Additive exPlanations (SHAP) and Gradient-weighted Class Activation Mapping (Grad-CAM) are employed for interpretability. Evaluated on the University of Texas at Dallas Multimodal Human Action Dataset (UTD-MHAD) dataset, TriNet-HAR achieves 98% accuracy, outperforming unimodal and bimodal baselines in precision, recall, and F1-score, demonstrating superior robustness, adaptability, and transparency.

Abstract Social-ecological systems (SES) are understood as human–environment systems that encompass all ecological and technological structures relevant to, and shaped by, human actions. They are complex systems in nature, comprising two main components – social and ecological – along with their numerous subcomponents. For analytical purposes, the SES concept must be both operational and measurable, which requires defining spatial and temporal boundaries. In this study, the municipality was selected as the basic territorial and administrative unit. In spite of their importance, evaluations and mapping of SES remain scarce, particularly in Latin America. This study assessed the SES status in a region of central-eastern Mexico - a priority biocultural region for conservation and development. The aim was to identify critical areas for interventions to enhance socio-ecological resilience and understand the key factors influencing SES. Results indicated a generally low SES status per municipality, with values ranging from -0.244 to 0.361. The highest values were observed in municipalities with high ecological integrity, low sensitivity to biophysical hazards, and stable ecosystem service availability. Spatial correlation analysis revealed a cold spot in the south-eastern area, where municipalities exhibited the lowest resilience and SES status. In contrast, a hotspot with the highest values was found in the centre of the study area, where the infrastructure (i.e., roads, hospitals) was first developed because of silver mining. A generalized linear model showed a negative correlation between Indigenous people and SES status. Urgent actions are needed to improve communication, healthcare, education, economic opportunities, to reduce their vulnerability to environmental challenges.

This study examines the emergence of problematic cannabis use as a dynamic, networked, and affective process. Drawing on Actor-Network Theory and Deleuze's concept of affect, the analysis explores how shifting assemblages of human and non-human actors shape agency over time. The data consist of interviews with 23 participants and 16 written accounts submitted by individuals who self-identified their cannabis use as problematic. Through close reading of participants' narratives combined with network mapping, three phases were identified: opening, sustaining, and restricting networks. In the opening phase, network relations expand affective capacity by opening up new ways of thinking and extending the body's range of movement in space. In the sustaining phase, these relations stabilize as cannabis becomes part of everyday routines that reproduce familiar rhythms rather than generating new openings. In the restricting phase, configurations involving specific material fixators (such as the bed, sofa, or digital devices) begin to limit spatial, emotional, and embodied possibilities, resulting in a contraction of affective capacity. These phases offer a heuristic framework for analyzing how problematic use stabilizes within socio-material assemblages, rather than as an outcome of isolated individual vulnerabilities. This perspective contributes to ongoing discussions in drug research by emphasizing relational dynamics and may inform treatment by identifying core networked actors that sustain or limit agency. Future research should examine how restricting networks may unravel and give rise to new affective landscapes.