Search Results (499)

Accurate detection and counting of wheat spikes are crucial for yield estimation and variety selection in precision agriculture. However, challenges such as complex field environments, morphological variations, and small target sizes hinder the performance of existing models in real-world applications. This study proposes FEWheat-YOLO, a lightweight and efficient detection framework optimized for deployment on agricultural edge devices. The architecture integrates four key modules: (1) FEMANet, a mixed aggregation feature enhancement network with Efficient Multi-scale Attention (EMA) for improved small-target representation; (2) BiAFA-FPN, a bidirectional asymmetric feature pyramid network for efficient multi-scale feature fusion; (3) ADown, an adaptive downsampling module that preserves structural details during resolution reduction; and (4) GSCDHead, a grouped shared convolution detection head for reduced parameters and computational cost. Evaluated on a hybrid dataset combining GWHD2021 and a self-collected field dataset, FEWheat-YOLO achieved a COCO-style AP of 51.11%, AP@50 of 89.8%, and AP scores of 18.1%, 50.5%, and 61.2% for small, medium, and large targets, respectively, with an average recall (AR) of 58.1%. In wheat spike counting tasks, the model achieved an R² of 0.941, MAE of 3.46, and RMSE of 6.25, demonstrating high counting accuracy and robustness. The proposed model requires only 0.67 M parameters, 5.3 GFLOPs, and 1.6 MB of storage, while achieving an inference speed of 54 FPS. Compared to YOLOv11n, FEWheat-YOLO improved AP@50, AP_s, AP_m, AP_l, and AR by 0.53%, 0.7%, 0.7%, 0.4%, and 0.3%, respectively, while reducing parameters by 74%, computation by 15.9%, and model size by 69.2%. These results indicate that FEWheat-YOLO provides an effective balance between detection accuracy, counting performance, and model efficiency, offering strong potential for real-time agricultural applications on resource-limited platforms. Full article

This study predicts sedimentary architectures and facies distribution within the Pliocene prograding prism of the Roussillon Basin (Gulf of Lion, France), developed along an onshore–offshore continuum. Boreholes and outcrops provide facies-scale observations onshore, while seismic data capture basin-scale structures offshore. Forward seismic modeling bridges spatial and scale gaps between these datasets, yielding characteristic synthetic seismic signatures for the sedimentary facies associations observed onshore, used as analogs for offshore deposits. These signatures are then identified in offshore seismic data, allowing seismic profiles to be populated with sedimentary facies without a well tie. Predicted offshore architectures are consistent with shoreline trajectories and facies successions observed onshore. The Roussillon prism records passive margin reconstruction in the Mediterranean Basin following the Messinian Salinity Crisis, through the following three successive depositional profiles marking the onset of infilling: (1) Gilbert deltas, (2) wave- and storm-reworked fan deltas, and (3) a wave-dominated delta. Offshore, transitions in clinoform type modify sedimentary architectures, influenced by inherited Messinian paleotopography. This autogenic control generates spatial variability in accommodation, driving changes in depositional style. Overall, this multi-scale and integrative approach provides a robust framework for predicting offshore sedimentary architectures and can be applied to other deltaic settings with limited land–sea data continuity. Full article

Multi-modal sensor fusion has become a cornerstone of robust autonomous driving systems, enabling perception models to integrate complementary cues from cameras, LiDARs, radars, and other modalities. This survey provides a structured overview of recent advances in deep learning-based fusion methods, categorizing them by architectural paradigms (e.g., BEV-centric fusion and cross-modal attention), learning strategies, and task adaptations. We highlight two dominant architectural trends: unified BEV representation and token-level cross-modal alignment, analyzing their design trade-offs and integration challenges. Furthermore, we review a wide range of applications, from object detection and semantic segmentation to behavior prediction and planning. Despite considerable progress, real-world deployment is hindered by issues such as spatio-temporal misalignment, domain shifts, and limited interpretability. We discuss how recent developments, such as diffusion models for generative fusion, Mamba-style recurrent architectures, and large vision–language models, may unlock future directions for scalable and trustworthy perception systems. Extensive comparisons, benchmark analyses, and design insights are provided to guide future research in this rapidly evolving field. Full article

Assessing teaching behavior is essential for improving instructional quality, particularly in Physical Education, where classroom interactions are fast-paced and complex. Traditional evaluation methods such as questionnaires, expert observations, and manual discourse analysis are often limited by subjectivity, high labor costs, and poor scalability. These challenges underscore the need for automated, objective tools to support pedagogical assessment. This study explores and compares the use of Transformer-based language models for the automatic classification of teaching behaviors from real classroom transcriptions. A dataset of over 1300 utterances was compiled and annotated according to the teaching styles proposed in the circumplex approach (Autonomy Support, Structure, Control, and Chaos), along with an additional category for messages in which no style could be identified (Unidentified Style). To address class imbalance and enhance linguistic variability, data augmentation techniques were applied. Eight pretrained BERT-based Transformer architectures were evaluated, including several pretraining strategies and architectural structures. BETO achieved the highest performance, with an accuracy of 0.78, a macro-averaged F1-score of 0.72, and a weighted F1-score of 0.77. It showed strength in identifying challenging utterances labeled as Chaos and Autonomy Support. Furthermore, other BERT-based models purely trained with a Spanish text corpus like DistilBERT also present competitive performance, achieving accuracy metrics over 0.73 and and F1-score of 0.68. These results demonstrate the potential of leveraging Transformer-based models for objective and scalable teacher behavior classification. The findings support the feasibility of leveraging pretrained language models to develop scalable, AI-driven systems for classroom behavior classification and pedagogical feedback. Full article

This paper introduces DKR-YOLO, an advanced deep learning framework designed to empower the digital preservation and sustainable management of ancient mural heritage. Building upon YOLOv8, DKR-YOLO integrates innovative components—including the DySnake Conv layer for refined feature extraction and an Adaptive Convolutional Kernel Warehouse to optimize representation—addressing challenges posed by intricate details, diverse artistic styles, and mural degradation. The network’s architecture further incorporates a Residual Feature Augmentation (RFA)-enhanced FPN (RE-FPN), prioritizing the most critical visual features and enhancing interpretability. Extensive experiments on mural datasets demonstrate that DKR-YOLO achieves a 43.6% reduction in FLOPs, a 3.7% increase in accuracy, and a 5.1% improvement in mAP compared to baseline models. This performance, combined with an emphasis on robustness and interpretability, supports more inclusive and accessible applications of AI for cultural institutions, thereby fostering broader participation and equity in digital heritage preservation. Full article

The Building Houses on Rented Land Model (BHRLM) was a pivotal land development model that drove Shanghai’s urbanization in the early modern era. This research examines the spatial generative mechanism of the Prepaid Building Houses on Rented Land Model (PBHRLM), prevalent during 1938–1941. It reveals how the wartime economic environment enabled interest alliances constituted with developers, landowners, and tenants to stimulate urban spatial growth. Firstly, we aim to analyze the features of architectural types linked to the PBHRLM using data-driven methods. Secondly, we aim to apply financial capital theory to investigate the innovations of financing methods. Finally, we draw on speculation theory to establish connections between the features of architectural types and the innovations of financing methods. The results include the following: (1) The PBHRLM’s dominant architectural types—new-styled lane houses, semi-shikumen lane houses, and garden houses—shared low-rise, high-density spatial features. (2) The PBHRLM’s innovations of financing methods lie in its convergence of financing and profitability, reflecting developers’ speculative intent. The research concludes that the PBHRLM operated as a spatial actuarial practice. Through risk games, the developers utilized the model to liberate land development from the control of financial capital and achieved multi-stakeholder synergy, generating small-scale, dispersed land development patterns. At the same time, surging housing demand thus perpetuated architectural types catering to the middle class with low-rise, low-tech tectonics and independent dwelling styles that continued to densely populate Shanghai concessions. Full article

The evolution of agricultural civilization is closely related to the social changes in ancient China, with Fuzhou being home to a large number of traditional granary buildings with distinctive regional characteristics. This study employs field surveys, a literature review, architectural mapping, and comparative analysis to explore whether there is mutual verification between the “room-style” granaries in Fuzhou and related agricultural imagery. The findings reveal that (1) the granary buildings in Fuzhou city generally follow the ancient raised-platform structure and are organically integrated with the local courtyard-style dwellings, forming a unique “room-style” granaries. Their layout and structure not only adapt to the local natural environment but also reflect the ancient craftsmen’s deep understanding of material properties and structural mechanics. (2) The spatial layout and functions of traditional granary buildings have evolved with social changes. Their development has been profoundly influenced by Zhuzi’s granary system and Neo-Confucian thought, gradually forming a hybrid space that combines storage and residential functions, integrating both practicality and esthetics. This evolutionary process not only reflects the flexibility and adaptability of the ancient storage system but also demonstrates the influence of social and cultural factors in shaping architectural space. Currently, there are the following gaps in the research on traditional granaries in Fuzhou City: a lack of analysis on the form and structural patterns of local granary buildings, insufficient cross-verification between documentary records and physical remains, and inadequate research on the construction wisdom of traditional granary buildings. This study provides valuable insights into the research of ancient architectural art, cultural exchange, and regional construction. Full article

Fine-grained modeling of driving styles is critical for decision making in autonomous driving. However, existing methods are constrained by the high cost of manual labeling and a lack of interpretability. This study proposes an unsupervised disentanglement framework based on a variational autoencoder (VAE), which, for the first time, enables the automatic extraction of interpretable driving style representations from car-following trajectories. The key innovations of this work are threefold: (1) a dual-decoder VAE architecture is designed, leveraging driver identity as a proxy label to guide the learning of the latent space; (2) self-dynamics and interaction dynamics features are decoupled, with an attention mechanism employed to quantify the influence of the lead vehicle; (3) a bidirectional interpretability verification framework is established between latent variables and trajectory behaviors. Evaluated on a car-following dataset comprising 25 drivers, the model achieves a Driver Identification accuracy of 98.88%. Mutual information analysis reveals the physical semantics encoded in major latent dimensions. For instance, latent dimension z₂₂ is strongly correlated with the minimum following distance and car-following efficiency. One-dimensional latent traversal further confirms that individual dimensions modulate specific behavioral traits: increasing z₂₂ improves safety margins by 18% but reduces efficiency by 23%, demonstrating that it encodes a trade-off between safety and efficiency. This work provides a controllable representation framework for driving style transfer in autonomous systems and offers a more granular approach for analyzing driver behavior in car-following scenarios, with potential for extension to broader driving contexts. Full article

Numerous well-preserved ancestral halls dating back to the Ming and Qing dynasties are found throughout Southern Anhui, China. Among the architectural elements in these ancestral halls, the “Xuan” plays a significant decorative role. Nevertheless, scholarly research on the “Xuan” in this region remains limited, particularly in the context of the temporal and regional evolution of its decorative art. This study examines the decorative characteristics and spatiotemporal evolution of the front Xuan of 115 ancestral halls across Southern Anhui belonging to the Ming and Qing periods. Using field surveys, a combination of qualitative and quantitative analyses, and ArcGIS kernel density analysis, this study identifies key trends in Xuan evolution. Specifically, the findings indicate that the Chuanpeng Xuan emerged as the predominant style, replacing the earlier Renzi Xuan after the mid-Ming period. By the late Qing period, the primary structural components of the Xuan showed increasing standardization, while accessory elements showed notable diversification. During the commercial boom of the mid-Qing era, flood-dragon carvings and S-shaped short beams became especially prevalent. Spatially, kernel density analysis demonstrated a core-periphery distribution pattern: mainstream styles were concentrated in central counties, whereas a greater variety of stylistic combinations and niche forms emerged in border regions. These evolutionary patterns reflect the broader sociohistorical dynamics of this period, including ritual reforms during the Ming dynasty, the patronage of Huizhou merchants, innovations in carving tools, and wartime resource constraints. These findings provide diagnostic criteria for dating unidentified ancestral halls and offer a practical reference for the conservation of architectural heritage. Full article

We introduce a code-only approach for modeling web page layouts directly from their source code (HTML and CSS only), bypassing rendering. Our method employs a neural architecture with specialized encoders for style rules, CSS selectors, and HTML attributes. These encodings are then aggregated in another neural network that integrates hierarchical context (sibling and ancestor information) to form rich representational vectors for each web page’s element. Using these vectors, our model predicts eight spatial relationships between pairs of elements, focusing on edge-based proximity in a multilabel classification setup. For scalable training, labels are automatically derived from the Document Object Model (DOM) data for each web page, but the model operates independently of the DOM during inference. During inference, the model does not use bounding boxes or any information found in the DOM; instead, it relies solely on the source code as input. This approach facilitates structure-aware visual analysis in a lightweight and fully code-based way. Our model demonstrates alignment with human judgment in the evaluation of web page similarity, suggesting that code-only layout modeling offers a promising direction for scalable, interpretable, and efficient web interface analysis. The evaluation metrics show our method yields similar performance despite relying on less information. Full article

Convolutional Neural Networks (CNNs) are a highly used machine learning architecture in various fields. Typical descriptions of CNNs are based on low-dimension and tensor representations in the feature extraction part. In this article, we extend the setting of CNNs to any arbitrary dimension and linearize the whole setting via the typical layers of neurons. In essence, a partial and a full network construct the entire process of a standard CNN, with the partial network being used to linearize the feature extraction. By doing so, we link the tensor-style representation of CNNs with the pure network representation. The outcomes serve two main purposes: to relate CNNs with other machine learning frameworks and to facilitate intuitive representations. Full article

This article presents an analysis of the architectural integration of two new buildings implemented on the Gdańsk University of Technology campus (Poland) as a case study of combining contemporary technologies with cultural continuity. The buildings, designed by the author of the article, who is the main designer, are a conscious response to the historical urban and architectural context of the campus, the development of which started at the beginning of the 20th century in the style of Dutch Neo-Renaissance. The new buildings refer to the architectural heritage of the university through their scale and colors, but their form, details and applied technological solutions clearly reflect modernity. A particularly important element of their modern character is the implementation of advanced pro-ecological systems for obtaining energy from renewable sources (RES), which fits into the current climate challenges and the role of the technical university as a promoter of sustainable development. The article discusses how architecture, materials and modern building systems were used to create a dialogue between tradition and innovation. The analysis is based on design documentation and planning conditions, and its background is a broader discourse on culturally sustainable architecture. Conscious of other, more conservative views, the author puts forward the thesis that cultural continuity does not require stylistic imitation, but conscious, contextual reinterpretation. The results of the article enrich the debate on the development of academic campuses, heritage-responsible design and the role of the architect in shaping a space that connects the future with the past. The main research contribution of the article is the presentation of an original method of designing architectural objects that integrates advanced pro-ecological technologies with a contextual reinterpretation of architectural heritage, which constitutes a new perspective in the discussion on culturally sustainable architecture. Full article

As the cathedral serving Zuoyun and parts of Inner Mongolia, the Batayizi Church in Datong, Shanxi is the largest surviving Italian Gothic-style Catholic church in the region. The church features a rigorous layout and refined details, making it a significant case study for the dissemination and development of Western architecture in China. Previous studies have focused on local chronicles, aesthetic analyses, and the indigenization of Catholic churches in Shanxi. Due to the scarcity of archival materials, research on the architecture itself has not yet been conducted. The article first summarizes the construction rules of local form and order of Italian Gothic churches based on related church remains and literature. Next, it establishes the architectural form of the church by combining construction rules and field surveys. Finally, the reconstruction design of the church is completed. As the first reconstruction study of the Batayizi Church, this paper attempts to explore a Reconstruction path based on the construction of local form and order of the church and systematically restores the main facade, floor plan, and structural form of the church. The results not only provide insights for the reconstruction of modern Catholic churches in Shanxi but also offer new ideas and methods for the study of the localization of Western architecture in China. Full article

Topophilia, understood as a form of relationship between humans and their environment, can manifest in diverse ways—not only across various domains of art and life but also within the realm of music. This article seeks to expand the thesis of topophilia as a category defining the musical space of creation, performance, and perception of a musical work. Topophilia is seen here in the context of human activity in the artistic dimensions—philosophical, creative, architectural, and environmental. The methodological background is derived from the philosophy of place, phenomenology of perception, and musical analysis. This provides the opportunity to apply hermeneutic–philosophical analysis with elements of the theory of place. The thesis of this study is probably one of the first approaches to the category of topophilia in musical analysis, examining the style of composers, such as J.S. Bach, F. Chopin, K. Szymanowski, W. Lutosławski, A. Webern, and I. Xenakis, enriched with elements of musical performance. Full article

This study, performed in an Iranian context, explores how specific architectural attributes of shopping centers can influence public preferences, with the aim of supporting the development of more sustainable and user-oriented urban environments. A discrete choice experiment involving 260 participants was conducted to assess preferences across seven architectural variables, each presented at varying levels: entrance position, openness (i.e., transparency through windows), architectural style, materials, window shape, scale, and symmetry. Participants evaluated paired facade images and selected their preferred designs, enabling an analysis of how these attributes impact consumer choices. The findings indicate that most variables significantly influenced facade preferences, except for arched windows and low levels of openness. In contrast, high openness emerged as the strongest positive predictor of preference. Participants also showed a marked preference for large-scale (inhumanly scaled) facade attributes, rectangular windows, extruded entrances, asymmetrical compositions, and concrete materials. Moderate preferences were observed for symmetrical designs, mixed window shapes, contemporary and postmodern styles, and brick materials. Conversely, neoclassical style, recessed entrances, stone material, and smaller-scale (humanly scaled) facades received the lowest preference ratings. These results might offer valuable insights for architects and urban planners and guide the creation of more attractive and functional shopping centers, ultimately enhancing the quality of urban life. Full article