Search Results (854)

Edge vision systems require convolutional neural networks (CNNs) that preserve recognition accuracy under strict storage, computation, and latency constraints. Although ResNet18 is a compact residual backbone, direct deployment on resource-constrained devices remains costly, whereas simple channel reduction weakens representation capacity. This study aims to build a deployable ResNet18-based classifier that reduces model complexity while recovering the accuracy lost during compression. We propose a lightweight framework that combines global channel scaling, a re-parameterized attention residual block, and teacher–student knowledge distillation. The proposed block uses multi-branch convolution and squeeze-and-excitation attention during training, then folds the linear branches into a single 3-by-3 convolution for inference. Experiments on CIFAR-100 show that the final model reduces parameters from 11.220 M to 2.841 M, retains comparable Top-1 accuracy (0.7579 vs. 0.7606), improves Top-5 accuracy (0.9340 vs. 0.9253), and reduces graphics processing unit (GPU) batch inference latency from 3.279 ms to 2.161 ms. Deployment on PYNQ-Z2 verifies the complete camera-based CPU-side inference workflow, with an average end-to-end latency of 421.467 ms/frame. The results indicate that residual topology preservation, re-parameterized feature enhancement, and distillation form a practical route for edge-oriented lightweight CNN deployment. Full article

This study investigates place-based barriers to initial teacher training (ITT) recruitment in rural-coastal regions of England, focusing on Cornwall as a case study. Utilizing semi-structured interviews with nine ITT provider leaders and nine trainee teachers, the research applies the concept of educational isolation to ITT providers in areas that are geographically remote, socioeconomic disadvantaged, and culturally isolated. The analysis is framed by the critical pedagogy of place and social capital theory, moving beyond deficit-based interpretations of rurality to critically examine how place-based inequities are produced through urban-normative policy and resource allocation. Primary data were analyzed using reflexive thematic analysis. Four substantive themes emerged: transport dependency and accessibility constraints that structurally exclude lower-income and disabled trainees; housing displacement driven by the tourist economy, which compounds financial insecurity; an “employment precarity problem” where localized primary school oversaturation coexists with secondary teacher shortages; and cultural and professional isolation that disproportionately impacts ethnically diverse trainees in demographically homogeneous communities. The research further identifies that community resilience, while enabling individuals to navigate structural barriers, can obscure infrastructural inadequacy and diminish impetus for systemic policy reform. This paper contributes to international scholarship on spatial justice and rural teacher education by presenting an integrated conceptual framework with transferable relevance to similar rural-coastal and peripheral contexts globally and by offering policy recommendations for place-weighted ITT funding, infrastructure investment in educationally isolated areas, and the development of collaborative provider models. Full article

The objective of this study is to explore the beliefs of teacher educators (TE) and pre-service teachers (PST) regarding cultural diversity and its relationship with teaching-learning practices, to contribute to strengthening Initial Teacher Education (ITE) with an intercultural approach. A quantitative observational, descriptive, and cross-sectional pilot study was conducted using a survey administered to 230 PST and TE from 18 Chilean universities. The results showed that TE exhibited higher levels of negative beliefs (d = 1.00; p < 0.001) and less favorable educational practices to cultural diversity (p = 0.012). Although no statistically significant differences were observed in favorable educational practices, a trend toward higher scores was identified among PST. Regression models indicated that the teaching role was significantly associated with negative beliefs, whereas the other variables showed no significant associations. Overall, the findings suggest a dissociation between stated and structural beliefs regarding cultural diversity, which may influence the implementation of intercultural pedagogical practices in teacher education. They underscore the need to strengthen intercultural training—particularly among TE—through critical reflection to align beliefs and pedagogical practices, and to advance future research on the evolution of these beliefs and their impact on teaching practice. Full article

Urban low-altitude unmanned aerial vehicle (UAV) logistics networks face critical operational bottlenecks due to complex three-dimensional spatial blockages, continuous communication diffraction, and severe vulnerability to information-layer threats such as Automatic Dependent Surveillance—Broadcast (ADS-B) signal anomalies. To address these interconnected challenges, this paper proposes an event-driven, cloud–edge collaborative digital twin framework to guarantee continuous multi-link communication and flight safety. The architecture operates through a dual-tier “Teacher–Student” paradigm. Under secure conditions, a cloud digital twin acts as a high-capacity “Teacher,” employing Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to partition heterogeneous user topologies. It then utilizes an energy-guided stochastic diffusion sampling (EGSDS) method to refine initial macroscopic routing, generating precise, outage-free global trajectories by systematically minimizing non-line-of-sight (NLoS) observation penalties and kinematic regularization costs. To counteract signal anomalies, a distributed Time Difference of Arrival (TDOA) anchor network continuously validates UAV coordinate integrity. If a threshold is breached, control authority is instantly transferred to the UAV’s edge digital twin. This resource-constrained edge tier relies on a localized “Student” network trained via progressive distillation. By compressing the computationally heavy iterative diffusion process into a rapid one-step inference model, the UAV autonomously generates a secure, short-range emergency path that strictly adheres to minimum communication thresholds. Once interference clears, the cloud seamlessly regains control to complete the logistics mission. Experimental results demonstrate that the proposed scheme significantly outperforms conventional heuristic routing methods in cloud-based scenarios. Furthermore, the edge-based distillation mechanism substantially improves the overall trajectory survival rate under signal anomalies, ensuring resilient and continuous logistics operations. Full article

Special education teachers often lack training to implement research-evaluated writing programs with fidelity, which contributes to insufficient instruction for students with disabilities. This study addresses a research gap: the limited documentation of implementation fidelity in French spelling interventions that integrate assistive technologies (ATs) for learners aged 9–12 with dysorthographia. Grounded in a theoretical foundation that coordinates alphabetic, orthographic, and morphographic processes within an explicit instruction sequence (explanation, modeling, guided practice, and independent application), the program aligned text-to-speech and word prediction with targeted spelling goals. Using a mixed-methods design, six elementary students participated in a single-case protocol with a transformative sequential design over 20 weeks. Four teachers received targeted training (theoretical + practical) and delivered explicit, individualized instruction during a 10-week intervention. Content analysis of teacher and researcher logs showed high, yet context-responsive, fidelity with variations by student profile, school context, and teacher. Converging quantitative and qualitative patterns suggest improvements in word-level accuracy/fluency and highlight training/coaching as a driver of fidelity. The discussion provides actionable implications for professional learning, school scheduling and dosage protection, and future research that multimodalizes fidelity evidence and instruments AT orchestration across the writing cycle. Full article

Digital transformation is reshaping technical and vocational education and training (TVET), yet the behavioral processes through which students’ digital learning competence is associated with learning performance remain underexplored. Drawing on Biggs’ presage–process–product (3P) model, this cross-sectional study examined a dual-path moderated mediation model in which active and rule-based learning participation served as differentiated process pathways, while teacher–student interaction and curriculum practicality were specified as contextual moderators. Survey data were collected from 3693 students in Chinese higher vocational colleges. Hierarchical regression and bootstrapped moderated mediation analyses indicated that digital learning competence was positively associated with learning performance. Active learning participation mediated this association, whereas rule-based learning participation did not function as a stable positive mediator. At higher levels of teacher–student interaction and curriculum practicality, digital learning competence showed stronger associations with active learning participation and stronger indirect associations with learning performance. The rule-based pathway appeared more conditional and reflected an externally regulated, prescribed-task-oriented form of behavioral participation, rather than a stable process pathway associated with deep learning. These findings extend the 3P model to digital learning in higher vocational education, differentiate behavioral forms of participation, and highlight the importance of interactive and practice-oriented instructional contexts. Full article

Generative AI tools are increasingly being adopted in education, where they function as collaborators that can provide feedback, suggest alternatives, and scaffold learning. In this paper, I conducted an autoethnographic study by examining my experience as a teacher-researcher integrating generative AI tools as a More Knowledgeable Other (MKO) within the context of game design education. Drawing on Vygotsky’s sociocultural theory, this study documents how generative AI can facilitate creative learning by extending learners’ capacity to ideate, iterate, and reflect on their design processes. This study further reflects on instructional practices and observations of learners engaging with AI-supported creative activities across workshops and training programs. My reflections reveal that generative AI tools enhance feedback loops, accelerate prototyping, and democratize access to mentorship by providing context-aware guidance. However, they also introduce challenges related to illusions of competence, a lack of internalization, and reduced iteration design depth. Future work will explore structured pedagogical models that balance human mentorship with AI-assisted guidance, aiming to establish ethical, adaptive, and creativity-centered frameworks for using generative AI in game design education. Through this lens, this study contributes to an emerging understanding of AI-enabled learning partnerships and their implications for cultivating innovation and talent in the creative industries. Full article

Berry thinning is a fundamental operation in modern vineyard management, and future robotic thinning systems have the potential to reduce labor intensity and improve operational consistency. However, automated berry thinning under field conditions is still constrained by insufficient berry-level segmentation accuracy, difficulty in recognizing occluded berries, and high missed-detection rates for small berries. These limitations mainly arise from dense berry arrangements, severe mutual occlusion, and the subtle visual features of small targets. To address these challenges, this study developed a lightweight grape berry instance segmentation and thinning decision-support method based on YOLOv8n-seg. A two-stage knowledge distillation strategy, using Mask R-CNN and YOLOv8l-seg as teacher models, was combined with 30% backbone pruning to improve the recognition of occluded and small berries while maintaining model efficiency. Subsequently, the DBSCAN clustering algorithm was used to analyze berry centroid coordinates and equivalent diameters extracted from instance segmentation masks, thereby generating preliminary thinning-target recommendations based on local berry density and berry size. The model was trained and evaluated on a self-constructed dataset containing 330 valid grape bunch images collected in 2025 from Yongming Vineyard, Lin’an District, Hangzhou, Zhejiang Province, China. The results showed that the optimized YOLOv8n-seg model achieved a box $mAP50\text{-}95$ of 0.8945 and a mask $mAP50\text{-}95$ of 0.7910, with an inference speed of 119.19 $FPS$ and 3.26 M parameters on an NVIDIA RTX 3060 Laptop GPU. Compared with the original YOLOv8n-seg model, the optimized model improved mask $mAP50\text{-}95$ by 1.20 percentage points, increased inference speed by 71.79 $FPS$, and reduced the number of parameters by 2.38 M. These results indicate that the proposed method improves grape berry instance segmentation performance while achieving a favorable balance among segmentation accuracy, lightweight characteristics, and inference efficiency. The proposed framework provides an offline RGB-based visual perception and preliminary thinning decision-support method for future grape berry thinning robots. However, because the current dataset was collected from Shine Muscat grape bunches at the berry enlargement stage in a single vineyard using the same imaging setup, the results should be interpreted as preliminary evidence under the specific cultivar, growth stage, vineyard, and imaging conditions of this study. Further validation across different grape cultivars, growth stages, vineyards, production seasons, camera systems, embedded platforms, and real robotic thinning operations is still required. Full article

Automated defect detection in precision agriculture serves as a critical technology for enhancing the quality of agricultural products. Although supervised-only semantic segmentation has demonstrated remarkable performance in citrus surface defect detection, it relies heavily on training with large-scale labeled data, which results in prohibitive acquisition costs. Semi-supervised learning mitigates reliance on labeled data by generating pseudo-labels. However, existing semi-supervised segmentation methods still face challenges. On the one hand, the instability of pseudo-labels and the propagation of noise can mislead the training of semi-supervised models. On the other hand, due to the lack of semantic constraints in feature learning, models often suffer from insufficient feature discriminability when handling complex samples, such as citrus surface defects characterized by similar textures and blurred boundaries. Therefore, this study proposes UP-ETS, a dual-guided semi-supervised semantic segmentation model based on the Mean Teacher–Student framework, specifically designed for the segmentation of complex citrus surface defects. UP-ETS employs Uncertainty Estimation (UE) based on Kullback–Leibler (KL) divergence to quantify the prediction discrepancy between the teacher and student models on blurred and ambiguous pixels. This mechanism guides the model to dynamically adjust weights, thereby reducing noise propagation and enhancing pseudo-label stability under complex citrus surface textures. Prototype Contrastive Learning (PCL) is utilized to align pixel-level features of difficult samples with class prototypes, optimizing the feature discriminability for complex citrus surfaces. Experimental results demonstrate that the UP-ETS model exhibits superior semi-supervised segmentation performance. Notably, at a labeled data ratio of only 1/16, the dice improved from 85.57% to 87.76% compared to the supervised-only baseline. Furthermore, the model shows significant performance enhancements in segmenting difficult samples, such as small targets, complex boundaries, and blurred regions. The results of ablation studies and t-SNE visualization prove the effectiveness of the proposed UE and PCL. These two methods synergistically guide the model to construct a feature space that is better structured and highly discriminative. Furthermore, UP-ETS outperforms various representative semi-supervised segmentation models in terms of segmentation performance, parameters, and inference speed. In cross-dataset validation, the model exhibits robust generalization capabilities, achieving performance comparable to supervised-only methods trained on the full augmented dataset. Consequently, the framework introduced in this study effectively mitigates the heavy dependency on annotated datasets, providing significant practical value for agricultural deployment. Full article

This study examines the perceptions of Palestinian Arab students enrolled in teacher training colleges regarding their religious, ethnic, and educational identities and investigates the associations between these identities and life orientation, self-mastery, and self-esteem. In addition, the study evaluates a hypothesized model in which life orientation is posited to mediate the relationship between religious identity, ethnic identity, self-esteem, self-mastery, and educational identity. The research includes a sample of 512 Arab Palestinian Israeli students studying in Israeli teacher training colleges who filled out an online questionnaire. The findings show that participants reported clear and coherent perceptions of their religious, ethnic, and educational identities, along with a generally positive life orientation and moderate levels of self-esteem and self-mastery. Significant correlations were found between the variables. Furthermore, religious and ethnic identity, self-esteem, self-mastery, and life orientation were all directly associated with educational identity. Bias-corrected bootstrap analyses confirmed significant indirect effects through positive life orientation, supporting the hypothesized mediation model. These findings help to illustrate the extent to which the religious and ethnic identities of indigenous multi-religious and multicultural minorities contribute to the growth and development of individuals and advocate for their strengthening, contrary to claims that belittle these identities or call for ignoring and suppressing them. Furthermore, the study underscores the potential role of identity awareness in fostering adaptive psychosocial adjustment and reducing social polarization. Full article

We propose an interactive approach to teaching Coulomb’s law and electrostatics in general, rooted in two complementary pedagogical methodologies: hyper-constructivism (H-C) and neo-realism. Unlike standard constructivism, our hyper-constructivist approach treats students’ prior ideas—even if incomplete or inconsistent—as essential “submerged logs” that teachers may use to guide students across the cognitive lake, toward the correct understanding. We implement a triadic model of cognitive didactics, balancing amusement (the ludic “hook”), formal teaching, and deepening scientific inquiry. Here, we present a hyper-constructivist path on electrostatics—Coulomb’s and Gauss’s laws. Through a sequential path of experiments involving plastic rods, “trained” aluminum cans, Volta’s electrophorus, and “Christmas” ornaments, we demonstrate how students can spontaneously formulate problems and bridge the gap between intuitive observations and complex effects of electrical polarization, going beyond the scholastic Coulomb’s law, via numerical modeling. The proposed interactive approach is rooted in phenomena-based learning and leverages discrepant events—surprising physical phenomena that challenge prior intuitions—as “ludic hooks” to trigger spontaneous inquiry and conceptual reconstruction. The main goal of our strategies is to trigger and develop young students’ interest in physics, which in many European countries is low. This method not only facilitates the acquisition of physical laws but also fosters “intellectual inquisitiveness” and social competencies, proving that well-rooted knowledge emerges from a synthesis of tangible experience and advanced scientific modeling. Our contribution constitutes a complex pedagogical proposal, iteratively developed and implemented in diverse didactical environments over several years. This paper presents a pedagogical proposal developed and refined through more than twenty years of educational practice. For teachers interested in implementing hyper-constructivist instruction, we provide a detailed teaching pathway on electrostatics, with didactical explanations and pedagogical notes. Full article

In this paper, we propose and evaluate a feature distillation technique for object detection under poor visibility conditions, and we analyze its impact when deployed on an FPGA platform. We demonstrate via extensive experiments how different detection architectures generalize across scenes, and we infer that a scale-permuted feature extraction is the ideal choice for detection tasks in unconstrained environments with an 11–12% gain. As verified by the experiments, image enhancement often fails to provide significant detection gains. We hence introduce a joint training in a scale-permuted student network that learns dehazed features from a dual teacher network without an explicit dehazing step. The student learns to replicate not only the teacher outputs but also the decision-making process of the teacher by using attention transfer. Although the overall goal is to produce a real-time system capable of providing driving assistance in challenging scenarios, the FPGA implementation of a scale-permuted network is the first of its kind. To achieve effective implementation of the model in FPGA technology, a high-level synthesis approach and model compression techniques are employed to obtain a deployment with a good trade-off between quality and memory footprint metrics. We develop two distilled models using the joint feature distillation technique and show that these perform better in poor visibility scenes when compared to other detectors with similar size or even bigger sizes in some cases. Our 8.5 M model shows an mAP gain of almost 1% compared to YOLOv10-M with 15 M parameters, on the Cityscapes Hazy dataset. On night images from the BDD dataset, our 8.5 M model shows an approximate mAP gain of 4% compared to YOLO26-S with 9.5 M parameters. We further perform cross-domain testing with the DriveIndia dataset to show that our models generalize well beyond the distillation distribution and can be used for generic driving scenarios. Full article

Accurate kidney segmentation in abdominal computed tomography (CT) images is important for quantitative analysis and computer-assisted clinical workflows, yet deploying deep learning models in practice remains challenging due to computational constraints. To address this, a disagreement-guided knowledge distillation (KD) framework is proposed for efficient kidney segmentation. The method introduces a spatial disagreement mask (Ω) to identify regions where teacher and student predictions diverge, enabling selective knowledge transfer focused on informative and error-prone areas while avoiding redundant supervision in well-predicted regions. In addition, a pixel-level annotated kidney segmentation dataset is created by extending a previously published abdominal CT dataset with kidney masks. The experimental results on both the in-house dataset and the KiTS19 benchmark show improved overlap-based segmentation performance, particularly in Dice and IoU, compared with supervised training and conventional pixel-wise KD. On the in-house dataset, Dice increases from 0.9239 to 0.9335 and IoU from 0.8720 to 0.8859, together with improved boundary-based distance metrics. On KiTS19, Dice improves from 0.8732 to 0.8812, primarily driven by improved kidney recall and a reduction in under-segmentation errors; however, boundary-based distance metrics remain more favorable for the conventional pixel-wise KD under domain shift. Additional experiments with a compact Attention U-Net-small student and stronger teacher sources further show that KD-Ω can improve compact student performance, although the magnitude of improvement depends on the teacher prediction profile. These findings indicate that the proposed framework provides an efficient and practical approach for enhancing lightweight segmentation models by prioritizing clinically relevant foreground preservation and reducing missed kidney regions under computational constraints. Full article

In contemporary society, dominated by visual communication and the intensive use of social media, educating learners to interpret images critically has become an essential task for all educational contexts. New generations live immersed in digital environments where images, often decontextualized or manipulated, contribute to shaping identities, relationships, and perceptions of reality. Within this scenario, image education should be understood not merely as technical literacy but as a critical and formative practice aimed at developing awareness, autonomous judgement, and analytical competence. The present contribution proposes a pedagogical reflection on the urgent need to integrate structured pathways of visual media literacy within school curricula, with particular attention to the role of schools as educational bastions in preventing phenomena such as the erosion of critical thinking or the diffusion of distorted and unrealistic visual models. This work situates itself within the interdisciplinary debate on Visual Education, highlighting the need to train teachers and educators capable of guiding children and adolescents in decoding visual messages and developing reflective thinking. Full article

Non-functional requirements (NFRs) are critical for ensuring software quality, yet they remain difficult to identify due to their implicit and loosely defined relationship with functional requirements (FRs). Existing research predominantly focuses on NFR classification, leaving the more practical problem of linking FRs with their corresponding NFRs largely underexplored. To bridge this gap, this research introduces Re-Distill, a framework that treats FR–NFR association as a retrieval task. It adopts a curriculum-guided, data-centric distillation strategy to improve semantic representations and capture the interdependencies between FRs and NFRs. The framework combines general semantic adaptation, domain-specific specialization, and teacher-guided hard-negative mining in a contrastive learning setting. During inference, it integrates dense and lexical retrieval with cross-encoder reranking to produce ranked NFR candidates for unseen FR queries. Experiments on an expanded FR–NFR dataset show consistent improvements throughout all training stages. The resulting model achieves a Recall@10 of 70.79%, significantly outperforming the zero-shot baseline (42.36% Recall@10). These results highlight the effectiveness of retrieval-based approaches for functional–non-functional requirement linking, providing a practical and scalable way to undertake software requirement analysis. Full article