Search Results (48)

Human cervical cancer and pre-cancer research relies on datasets scattered across modality-specific archives, imaging repositories, benchmark platforms, trial registries, and controlled-access catalogs. This fragmentation—combined with heterogeneous metadata, ambiguous use of “cervical” terminology, and inconsistent indexing of pre-cancer and screening/triage resources—limits reproducible discovery, access planning, and cross-modal benchmarking. We present the Cervical Cancer Dataset Catalog (CCDCAT), a machine-readable, versioned dataset of datasets that enumerates host-specific dataset-instance records anchored to stable identifiers and resolvable landing records within an explicitly declared discoverable source universe (U_v1.0) and a frozen discovery/labeling lexicon (Q_v1.0). The CCDCAT spans invasive cervical cancer, pre-cancer/dysplasia, and cervix-focused screening and triage phenotypes, and it covers molecular omics, imaging and microscopy (including cervix photography, cytology, and digital pathology), trial registry records, benchmark resources, and controlled-access catalogs represented as metadata with explicit access pathways. Eligibility and labels are assigned conservatively from source-provided metadata; when evidence is insufficient, the CCDCAT abstains rather than infers. In the initial release (CCDCAT-U_v1.0; v0.1), we enumerate 14 eligible dataset instances across 11 host systems within a declared universe of 21 sources. Releases include manuscript-ready tables and interoperable artifacts (schema, controlled vocabularies, provenance logs, abstention ledgers, and a queryable database), enabling reproducible filtering, linkage, and auditable reuse planning. Full article

Task planning for robots in real-life settings is inherently complex due to challenges such as identifying grounded sequences of actions to achieve a goal, bridging the gap between high-level planning and low-level execution, and addressing the computational constraints of robotic hardware. Additionally, an essential, but frequently underestimated, aspect is the robot’s ability to recognize its own limitations and effectively delegate tasks beyond its abilities to human collaborators. Recent advancements in integrating Large Language Models (LLMs) into robotics have increased the likelihood of generating unfeasible plans, making this capability even more critical for any robot that needs to interact with a real environment. To address these challenges, this paper presents a framework that integrates open-vocabulary online grounding, planning, and embodiment, emphasizing self-awareness and self-analysis. The proposed self-awareness model enables robots to assess task complexity and map task requirements to their own capabilities. This can be used in several ways, including strategically requesting human assistance when needed. By leveraging pre-trained foundation models, our framework supports a multi-role mechanism that enhances adaptability and ensures effective collaboration between robots and humans in real-world scenarios. The experiments were conducted using a TIAGo robot with the framework’s ability to identify and delegate challenging tasks on a set of nine real-life scenarios, spanning different complexity levels. The results show a significant improvement in task success rate, achieving 89.1% compared to 37.0% of the baseline, demonstrating the effectiveness of incorporating self-awareness into the planning process. The code and the additional materials have been publicly released on the project website. Full article

This article examines how kindness can be understood, expressed and enacted through intercultural citizenship education in higher education, with particular attention to equity, diversity and inclusion (EDI). Situated within a theoretical framework that brings together intercultural citizenship and EDI, the study argues that these fields are mutually reinforcing and that their integration is enriched by foregrounding kindness. Empirically, the article reports on a qualitative multiple case study conducted in 2023, involving university students from Argentina and the United Kingdom who collaboratively designed English language teaching materials focused on kindness. Data consisted of student-generated textual and artistic artefacts, including lesson plans, teachers’ notes, drawings, comics and other teaching materials, which were analysed using a multimodal approach. Across cases, kindness functioned as a relational disposition, ethical compass, emotional anchor and intentional action, serving as a pedagogical response to issues of gender inequality, mental health and disability inclusion. The study argues that a structured grammar of intercultural kindness offers a vocabulary that makes visible the relational, ethical, emotional and action-oriented dimensions through which kindness shapes the pedagogical enactment of intercultural citizenship and EDI. This approach demonstrates that kindness can be taught; however, its transformative potential depends on a deliberate political orientation. Full article

Natural-language-driven robot navigation in complex indoor environments requires the joint capability of high-fidelity scene representation, structured semantic reasoning, and executable path planning. To address this challenge, this paper proposes HOSG-Nav, a unified framework for natural-language-driven global navigation that integrates open-vocabulary 3D Gaussian scene representation, hierarchical semantic scene graph construction, and large-language-model-driven planning. First, an open-vocabulary 3D Gaussian field is constructed to jointly encode scene geometry, appearance, and semantic information, where compressed CLIP features are lifted into continuous 3D space and depth supervision is introduced to enhance geometric stability and metric-scale consistency. Second, the optimized Gaussian primitives are further abstracted into a semantic scene graph with a region–object hierarchical structure and traversable topological relations to support structured environment understanding. Finally, for natural language instructions, hierarchical semantic parsing is performed with the assistance of a large language model, and executable global navigation paths are generated through cross-modal target retrieval and graph-search-based planning. Experimental results on the Replica dataset demonstrate that HOSG-Nav achieves competitive performance in scene representation, semantic target retrieval, and global navigation, validating the effectiveness of jointly integrating multimodal 3D representation, hierarchical semantic abstraction, and language-guided planning. Full article

Visually impaired users face significant challenges in navigating complex indoor environments due to limited spatial awareness and lack of real-time semantic guidance. This paper proposes a multimodal navigation system integrating environmental perception with vision-language models (VLMs). It provides context-aware and explainable guidance without requiring additional infrastructure. The proposed system combines RTAB-Map for localization, YOLO-World for open-vocabulary object detection, and a lightweight language model for semantic reasoning and natural language interaction. To evaluate our system, experiments are conducted using the RePOPE benchmark to assess hallucination in vision-language understanding. Real-world indoor navigation experiments are also performed. The results show that integrating perception with language-based reasoning improves precision by up to 2.29% and consistently enhances F1-score compared to baseline VLM approaches. Real-world experiments further demonstrate reliable navigation performance, including multi-floor path planning and obstacle-aware guidance. Hence, the proposed system effectively enhances spatial understanding and reduces hallucination, providing a practical and scalable solution for assistive navigation. Full article

Benchmark-driven evaluation helps distinguish between planning quality and interface reliability when large language models are utilized for embodied reasoning in simulation. Our submission to the Embodied Agent Interface Challenge (EAI) is evaluated across four stages of the pipeline. These being goal interpretation, subgoal decomposition, action sequencing, and transition modeling. The tasks run in the BEHAVIOR and VirtualHome simulators, which use constrained action vocabularies, fixed-object inventories and symbolic state representations within a standard evaluation protocol. Our system accesses the OpenAI API using GPT-4.1 for BEHAVIOR, GPT-4.1-mini for VirtualHome, and GPT-5-mini in later exploratory experiments across both environments. The schemas for each task determine how the outputs are structured, and outputs are regenerated when they do not follow the specification. On the final public leaderboard, our system ranked eighteenth overall with a score of 57.92, achieving 68.88 on BEHAVIOR and 46.96 on VirtualHome. In this paper, we describe our approach and discuss what these observations suggest about the strengths and limitations of current language models when used for embodied reasoning. Full article

Object counting in remote sensing images is valuable for applications such as urban planning and environmental monitoring. However, it remains challenging due to heterogeneous annotations, semantic ambiguity in open-vocabulary queries, and performance degradation of small targets. To address these limitations, we propose DR-CLIP (Deformable Remote CLIP), a vision–language model for remote sensing image counting that incorporates deformable visual feature extraction with text-guided prediction. DR-CLIP includes a (1) Region-to-Instruction (R2I) mechanism to convert points, bounding boxes, and polygons into a unified image–text training representation, a (2) Multi-scale Deformable Attention (MSDA) to enhance discriminative feature extraction across extreme scale variations and cluttered backgrounds, and a (3) Text-Guided Counting Head that establishes robust cross-modal alignment through contrastive learning, achieving open-vocabulary counting capability without category-specific retraining. On DOTA-v2.0, DR-CLIP achieves a Mean Absolute Error (MAE) of 2.34 and a Root Mean Squared Error (RMSE) of 3.89, outperforming baselines by 19.0% in MAE. The MSDA module significantly increases Small-Object Recall (SOR) to 0.824, which is especially effective in situations involving dense and small object counting. In cross-modal retrieval, DR-CLIP attains R@1 scores of 68.3% (image-to-text) and 72.1% (text-to-image) on the Remote Sensing Image Captioning Dataset (RSICD). The framework generalizes robustly, with only 8.7% performance degradation in cross-domain tests, which is significantly lower than the 23.4% drop observed in baseline methods. Full article

This paper introduces a multi-turn large language model (LLM) jailbreaking attack called Defamiliarization, in which malicious queries are embedded within ostensibly harmless narratives. By reframing requests in “unmarked” contexts, LLMs can be coerced into producing undesirable outputs. A range of scenarios is documented, from planning ethically dubious actions to selectively overlooking critical events in literary texts, thereby exposing the limitations of alignment strategies predicated on detecting trigger words or semantic cues. Rather than substituting vocabulary, defamiliarization manipulates context and presentation, highlighting vulnerabilities that cannot be addressed by token-level fixes alone. Beyond demonstrating the effectiveness of defamiliarization as an attack strategy, evidence is presented of a systematic relationship between model scale and susceptibility. Experiments reveal that smaller-parameter models are significantly easier to manipulate using defamiliarized prompts. This finding raises important concerns regarding the growing popularity of lightweight, locally hosted LLMs, which are favored for their lower computational requirements but may lack alignment safeguards. A more holistic approach to LLM safety is advocated—one that incorporates insights from literary theory, ethics, and user experience—treating these models as interpretive agents. By doing so, defenses against covert manipulations can be strengthened and AI systems can remain aligned with human values. Full article

Semantic point cloud maps play a pivotal role in smart agriculture. They provide not only core three-dimensional data for orchard management but also empower robots with environmental perception, enabling safer and more efficient navigation and planning. However, traditional point cloud maps primarily model surrounding obstacles from a geometric perspective, failing to capture distinctions and characteristics between individual obstacles. In contrast, semantic maps encompass semantic information and even topological relationships among objects in the environment. Furthermore, existing semantic map construction methods are predominantly vision-based, making them ill-suited to handle rapid lighting changes in agricultural settings that can cause positioning failures. Therefore, this paper proposes a positioning and semantic map reconstruction method tailored for orchards. It integrates visual, LiDAR, and inertial sensors to obtain high-precision pose and point cloud maps. By combining open-vocabulary detection and semantic segmentation models, it projects two-dimensional detected semantic information onto the three-dimensional point cloud, ultimately generating a point cloud map enriched with semantic information. The resulting 2D occupancy grid map is utilized for robotic motion planning. Experimental results demonstrate that on a custom dataset, the proposed method achieves 74.33% mIoU for semantic segmentation accuracy, 12.4% relative error for fruit recall rate, and 0.038803 m mean translation error for localization. The deployed semantic segmentation network Fast-SAM achieves a processing speed of 13.36 ms per frame. These results demonstrate that the proposed method combines high accuracy with real-time performance in semantic map reconstruction. This exploratory work provides theoretical and technical references for future research on more precise localization and more complete semantic mapping, offering broad application prospects and providing key technological support for intelligent agriculture. Full article

Aims: To ask UK women with diabetes whether they have discussed sexual health with healthcare professionals (HCPs) during diabetes care, and to explore communication barriers. Methods: An online questionnaire was developed, based on a published HCP communication survey, piloted by six women with diabetes. A total of 163 participants, recruited via social media and HCP network, completed Part 1 by selecting Likert or narrative response options, providing descriptive data. We report proportions with 95% confidence intervals (Wilson); percentages are calculated using the number responding to each item. Item-level missingness is retained as a non-analysed category, and the n is reported per question. No inferential comparisons were planned a priori. After Part 1 completion, participants could choose to finish, or to continue to Part 2 questions regarding vulval anatomy, function, and vocabulary (77 completed 2A: 80 completed 2B). Part 2 data was analysed thematically. Results: During diabetes care, a minority of participants, 44/163 (27%), said they had ever discussed sexual health, or had been advised how to access sexual health support, 28/163 (17%). If an HCP discussed sexual health, many women said they expected to feel surprised, 114/163 (70%), or pleased, 88/163 (54%). Some participants said they expected HCPs would find the topic inappropriate, 56/163 (36%), or annoying, 44/163 (27%). Some participants expressed HCP gender preference (75/163 [46%] female and 4/163 [3%] male) for such discussion. Part 2 findings revealed unmet sexual health literacy needs with potential to impact on communication with HCPs. Conclusions: Women reported infrequent communication about sexual health and diabetes during diabetes care. Findings highlight potential communication barriers for some participants including the following: unmet educational needs regarding diabetes and sexual health, lack of confidence about available support, fear of a negative HCP response, and preference for the gender of the HCP. Whereas in previous research, HCPs feared upsetting women by discussing sexual health, many participants said they expected to respond positively. Full article

We present a bibliometric analysis of unmanned aerial vehicle (UAV) research that replaces simple keyword filtering with a context-aware, two-tier pipeline. Records from Web of Science and Scopus (198,152 total) were harmonized and de-duplicated in three stages (DOI, normalized title, blockwise Jaro–Winkler), yielding 129,124 unique items. To separate UAV work from entomology using overlapping vocabulary (e.g., swarm), we first applied rule-based weak labels with explicit UAV and insect regex families and a UAV context rule for “swarm,” then trained an elastic-net logistic regression on TF–IDF features and tuned the decision threshold to meet a high-precision target on a held-out split. The final corpus comprises 129,099 UAV records. Beyond lexical inventories, a keyword co-occurrence timeline shows reinforcement learning increasingly aligned with path planning and collision avoidance, while constraints such as energy and communication persist. A co-authorship network reveals bridging authors that connect guidance/control, perception, and communication subfields. The results show how UAV research is organized around central scientific problems and identify persistent obstacles such as energy efficiency, communication reliability, and robust decision-making in dynamic conditions. Full article

3D Multi-Object Tracking (3D MOT) is a critical task in autonomous systems, where accurate and robust tracking of multiple objects in dynamic environments is essential. Traditional approaches primarily rely on visual or geometric features, often neglecting the rich semantic information available in textual modalities. In this paper, we propose Text-Guided 3D Multi-Object Tracking (TG3MOT), a novel framework that incorporates Vision-Language Models (VLMs) into the YONTD architecture to improve 3D MOT performance. Our framework leverages RegionCLIP, a multimodal open-vocabulary detector, to achieve fine-grained alignment between image regions and textual concepts, enabling the incorporation of semantic information into the tracking process. To address challenges such as occlusion, blurring, and ambiguous object appearances, we introduce the Target Semantic Matching Module (TSM), which quantifies the uncertainty of semantic alignment and filters out unreliable regions. Additionally, we propose the 3D Feature Exponential Moving Average Module (3D F-EMA) to incorporate temporal information, improving robustness in noisy or occluded scenarios. Furthermore, the Gaussian Confidence Fusion Module (GCF) is introduced to weight historical trajectory confidences based on temporal proximity, enhancing the accuracy of trajectory management. We evaluate our framework on the KITTI dataset and compare it with the YONTD baseline. Extensive experiments demonstrate that although the overall HOTA gain of TG3MOT is modest (+0.64%), our method achieves substantial improvements in association accuracy (+0.83%) and significantly reduces ID switches (−16.7%). These improvements are particularly valuable in real-world autonomous driving scenarios, where maintaining consistent trajectories under occlusion and ambiguous appearances is crucial for downstream tasks such as trajectory prediction and motion planning. The code will be made publicly available. Full article

Urban flooding presents a growing challenge amid rapid urbanization, climate variability, and fragmented governance. Although simulation and risk assessment tools have advanced considerably, their integration into urban planning remains limited. This study utilized a comprehensive bibliometric analysis of 1293 articles from the Scopus database, selected through a PRISMA-guided workflow, to examine the temporal, structural, and conceptual evolution of simulation, flood risk, and planning in urban flood research from 1980 to 2025. The findings reveal a thematic progression from engineering-centric approaches to broader discourses on resilience, adaptation, and systemic risk. However, disciplinary fragmentation persists, with technical modeling, infrastructure planning, and governance still weakly connected. Despite a shared vocabulary around climate risk and resilience, practical integration into decision-making frameworks remains underdeveloped. The study highlights the need for more cohesive research-practice linkages and calls for frameworks that better align simulation outputs with urban planning imperatives. Full article

Background: Microsurgery is a highly complex and technically demanding field within reconstructive surgery, with outcomes heavily dependent on meticulous planning, precision, and postoperative monitoring. Over the last five years, artificial intelligence (AI) has emerged as a transformative tool across all phases of microsurgical care, offering new capabilities in imaging analysis, intraoperative decision support, and outcome prediction. Methods: A comprehensive narrative review was conducted to evaluate the peer-reviewed literature published between 2020 and May 2025. Multiple databases, including PubMed, Embase, Cochrane, Scopus, and Web of Science, were searched using combinations of controlled vocabulary and free-text terms relating to AI and microsurgery. Studies were included if they described AI applications during the preoperative, intraoperative, or postoperative phases of microsurgical care in human subjects. Discussion: Using predictive models, AI demonstrated significant utility in preoperative planning through automated perforator mapping, flap design, and individualised risk stratification. AI-enhanced augmented reality and perfusion analysis tools improved precision intraoperatively, while innovative robotic platforms and intraoperative advisors showed early promise. Postoperatively, mobile-based deep learning applications enabled continuous flap monitoring with sensitivities exceeding 90%, and AI models accurately predicted surgical site infections, transfusion needs, and long-term outcomes. Despite these advances, most studies relied on retrospective single-centre data, and large-scale, prospective validation remains limited. Conclusions: AI is poised to enhance microsurgical precision, safety, and efficiency. However, its integration is challenged by data heterogeneity, generalisability concerns, and the need for human oversight in nuanced clinical scenarios. Standardised data collection and multicentre collaboration are vital for robust, equitable AI deployment. With careful validation and implementation, AI holds the potential to redefine microsurgical workflows and improve patient outcomes across diverse clinical settings. Full article

Current policy recommendations for initial teacher education encourage teaching code-related literacy (phonics, phonological awareness, and phonemic awareness) over pedagogical knowledge, and engaging practice in learning to read. To enhance early childhood pre-service teacher (PST) practices, this mixed-methods pilot study investigated a tool to support PSTs studying birth-to-eight years teaching, pedagogical practice, and knowledge to teach code-related literacy and supplementary vocabulary in conjunction with quality children’s literature. The Non-Scripted Intentional Teaching (N-SIT) tool was developed and then trialled with early childhood PSTs (n = 24) in Queensland, Australia. The participants planned phonics learning experiences using the N-SIT and picture books (e.g., Pig the Pug; Snail and the Whale). Survey data gathered participants’ code-related literacy knowledge before and after the N-SIT training. The data revealed most PSTs felt well-to-somewhat prepared to teach beginning reading and vocabulary and less-to-somewhat prepared to teach phonics. The data further revealed that all participants could define phonics but reported mixed conceptual understandings of phonological and phonemic awareness. The PSTs’ knowledge of phonological awareness, phonemic awareness, and planning for phonics-focused teaching through children’s literature improved post-N-SIT activity. Planned direct systematic phonics instruction strategies through the intentional shared reading of children’s literature and the potential benefits of the N-SIT tool in early childhood initial teacher education are discussed. Full article