Search Results (1,374)

Accurate segmentation of retinal lesions is critical for the diagnosis and management of ophthalmic diseases, but pixel-level annotation is labor-intensive and demanding in clinical scenarios. To address this, we introduce a promptable segmentation approach based on prototype learning that enables precise retinal lesion segmentation from low-cost, coarse annotations. Our framework treats clinician-provided coarse masks (such as ellipses) as prompts to guide the extraction and refinement of lesion and background feature prototypes. A lightweight U-Net backbone fuses image content with spatial priors, while a superpixel-guided prototype weighting module is employed to mitigate background interference within coarse prompts. We simulate coarse prompts from fine-grained masks to train the model, and extensively validate our method across three datasets (IDRiD, DDR, and a private clinical set) with a range of annotation coarseness levels. Experimental results demonstrate that our prototype-based model significantly outperforms fully supervised and non-prototypical promptable baselines, achieving more accurate and robust segmentation, particularly for challenging and variable lesions. The approach exhibits excellent adaptability to unseen data distributions and lesion types, maintaining stable performance even under highly coarse prompts. This work highlights the potential of prompt-driven, prototype-based solutions for efficient and reliable medical image segmentation in practical clinical settings. Full article

Synthetic Aperture Radar (SAR) plays a pivotal role in remote-sensing target detection. However, domain shift caused by distribution discrepancies across sensors, coupled with the scarcity of target-domain samples, severely restricts the generalization and practical performance of SAR detectors. To address these challenges, this paper proposes a few-shot SAR target-detection framework tailored for cross-sensor scenarios (CS-FSDet), enabling efficient transfer of source-domain knowledge to the target domain. First, to mitigate inter-domain feature-distribution mismatch, we introduce a Multi-scale Uncertainty-aware Bayesian Distribution Alignment (MUBDA) strategy. By modeling features as Gaussian distributions with uncertainty and performing dynamic weighting based on uncertainty, MUBDA achieves fine-grained distribution-level alignment of SAR features under different resolutions. Furthermore, we design an Adaptive Cross-domain Interactive Coordinate Attention (ACICA) module that computes cross-domain spatial-attention similarity and learns interaction weights adaptively, thereby suppressing domain-specific interference and enhancing the expressiveness of domain-shared target features. Extensive experiments on two cross-sensor few-shot detection tasks, HRSID→SSDD and SSDD→HRSID, demonstrate that the proposed method consistently surpasses state-of-the-art approaches in mean Average Precision (mAP) under 1-shot to 10-shot settings. Full article

Accurate and automated detection of diseases is crucial for sustainable strawberry production. However, the challenges posed by small size, mutual occlusion, and high intra-class variance of symptoms in complex agricultural environments make this difficult. Mainstream deep learning detectors often do not perform well under these demanding conditions. We propose a novel detection framework designed for superior accuracy and robustness to address this critical gap. Our framework introduces four key innovations: First, we propose a novel attention-driven detection head featuring our Parallel Pyramid Attention (PPA) module. Inspired by pyramid attention principles, our module’s unique parallel multi-branch architecture is designed to overcome the limitations of serial processing. It simultaneously integrates global, local, and serial features to generate a fine-grained attention map, significantly improving the model’s focus on targets of varying scales. Second, we enhance the core feature fusion blocks by integrating Monte Carlo Attention (MCAttn), effectively empowering the model to recognize targets across diverse scales. Third, to improve the feature representation capacity of the backbone without increasing the parametric overhead, we replace standard convolutions with Frequency-Dynamic Convolutions (FDConv). This approach constructs highly diverse kernels in the frequency domain. Finally, we employ the Scale-Decoupled Loss function to optimize training dynamics. By adaptively re-weighting the localization and scale losses based on target size, we stabilize the training process and improve the Precision of bounding box regression for small objects. Extensive experiments on a challenging dataset related to strawberry diseases demonstrate that our proposed model achieves a mean Average Precision (MAP) of 81.1%. This represents an improvement of 2.1% over the strong YOLOv12-n baseline, highlighting its practical value as an effective tool for intelligent disease protection. Full article

The localization and pose estimation of clinicians in the operating room is a critical component for building intelligent perception systems, playing a vital role in enhancing surgical standardization and safety. Multi-view, multi-person 3D pose estimation is a highly challenging task—especially in the operating room, where the presence of sterile clothing, occlusion from surgical instruments, and limited data availability due to privacy concerns exacerbate the difficulty. While voxel-based 3D pose estimation methods have shown promising results in general scenarios, their performance is significantly challenged in surgical environments with limited camera views and severe occlusions. To address these issues, this paper proposes a fine-grained voxel feature reconstruction method enhanced with depth information, effectively mitigating projection errors caused by reduced viewpoints. Additionally, an attention mechanism is integrated into the encoder–decoder architecture to improve the network’s capacity for global information modeling and enhance the accuracy of keypoint regression. Experiments conducted in real-world operating room scenarios, using the Multi-View Operating Room (MVOR) dataset, demonstrate that the proposed method maintains high accuracy even under limited camera views and outperforms existing state-of-the-art multi-view 3D pose estimation approaches. This work provides a novel and efficient solution for human pose estimation (HPE) in complex medical environments. Full article

This study introduces a novel, two-stage extraction system that combines Ohmic-Accelerated Steam Distillation (OASD) with Supercritical CO₂ Extraction (SSCO₂) to efficiently recover bioactive compounds from plant-based wastes with varying cell wall complexities. Brewer’s spent grain (BSG) and cocoa shell were selected as representative models for soft and rigid cell wall structures, respectively. The optimized extraction process demonstrated significantly enhanced efficiency compared to traditional methods, achieving recovery rates in BSG of 89% for antioxidants, 91% for phenolic acids, and 90% for polyphenolic compounds. Notably, high yields of p-coumaric acid (95%), gallic acid (94%), ferulic acid (82%), quercetin (87%), and resveratrol (82%) were obtained with minimal cellular structural damage. For cocoa shells, despite their lignin-rich, rigid cell walls, recovery rates reached 73% for antioxidants, 79% for phenolic acids, and 74% for polyphenolic compounds, including chlorogenic acid (94%), catechin (83%), vanillin (81%), and gallic acid (94%). Overall, this hybrid technique significantly improved extraction efficiency by approximately 60% for BSG and 50% for cocoa shell relative to conventional approaches, highlighting its novelty, scalability, and potential for broad application in the sustainable valorization of diverse plant-based waste streams. This research presents a green and efficient platform suitable for valorizing agri-food by-products, supporting circular economy goals. Further studies may explore scale-up strategies and economic feasibility for industrial adoption. Full article

While irrigation is generally required for most summer crops in the Mediterranean region, increasingly scarce water supplies are leading to a demand for more efficient irrigation infrastructure. Here, we assess how three irrigation volumes—100 mm/week (simulating excess water), 55 mm twice per week (moderate supply), and a variable amount adjusted on a weekly basis according to crop water demand (AMP) applied once or twice weekly via drip irrigation—impacted the growth, yield, and ear traits of a local maize variety under low-input farming in central Portugal. We found that irrigation management significantly influenced grain yield and irrigation water use efficiency (IWUE), with the 55 mm treatment applied twice weekly achieving the highest yield (3504 kg ha⁻¹) and IWUE (7.2 kg ha⁻¹ mm⁻¹). The highest irrigation treatment (100 mm/weekly) impaired yield (996 kg ha⁻¹ and 1973 kg ha⁻¹, when water was applied in one or two events), likely due to nutrient leaching, and resulted in the lowest IWRU (1.2 kg ha⁻¹ mm⁻¹ and 2.5 kg ha⁻¹ mm⁻¹, respectively). Biweekly applications tended to increase crop height. Irrigation rate and frequency significantly affected kernel number and size, but not total ear weight or cob-to-ear weight ratio. These findings highlight the importance of irrigation frequency based on crop water demand over blanket approaches based on volume alone. Full article

With the globalization of trade, maritime transport is playing an increasingly strategic role in sustaining international commerce. As a result, research into the tracking and fusion of multi-source vessel data in canal environments has become critical for enhancing maritime situational awareness. In the existing research and development, the heterogeneity of and variability in vessel flow data often lead to multiple issues in tracking algorithms, as well as in subsequent trajectory-matching processes. The existing tracking and matching frameworks typically suffer from three major limitations: insufficient capacity to extract fine-grained features from multi-source data; difficulty in balancing global context with local dynamics during multi-scale feature tracking; and an inadequate ability to model long-range temporal dependencies in trajectory matching. To address these challenges, this study proposes the Shape Similarity and Generalized Distance Adjustment (SSGDA) framework, a novel vessel trajectory-matching approach designed to track and associate multi-source heterogeneous vessel data in complex canal environments. The primary contributions of this work are summarized as follows: (1) an enhanced optimization strategy for trajectory fusion based on Enhanced Particle Swarm Optimization (E-PSO) designed for the proposed trajectory-matching framework; (2) the proposal of a trajectory similarity measurement method utilizing a distance-based reward–penalty mechanism, followed by empirical validation using the publicly available FVessel dataset. Comprehensive aggregation and analysis of the experimental results demonstrate that the proposed SSGDA method achieved a matching precision of 96.30%, outperforming all comparative approaches. Additionally, the proposed method reduced the mean-squared error between trajectory points by 97.82 pixel units. These findings further highlight the strong research potential and practical applicability of the proposed framework in real-world canal scenarios. Full article

Graph convolutional networks (GCNs) have demonstrated remarkable effectiveness in processing graph-structured data and have been widely adopted across various domains. Existing methods mitigate over-smoothing through selective aggregation strategies such as attention mechanisms, edge dropout, and neighbor sampling. While some approaches incorporate global structural context, they often underexplore category-aware representations and inter-category differences, which are crucial for enhancing node discriminability. To address these limitations, a novel framework, CA-NodeNet, is proposed for semi-supervised node classification. CA-NodeNet comprises three key components: (1) coarse-grained node feature learning, (2) category-decoupled multi-branch attention, and (3) inter-category difference feature learning. Initially, a GCN-based encoder is employed to aggregate neighborhood information and learn coarse-grained representations. Subsequently, the category-decoupled multi-branch attention module employs a hierarchical multi-branch architecture, in which each branch incorporates category-specific attention mechanisms to project coarse-grained features into disentangled semantic subspaces. Furthermore, a layer-wise intermediate supervision strategy is adopted to facilitate the learning of discriminative category-specific features within each branch. To further enhance node feature discriminability, we introduce an inter-category difference feature learning module. This module first encodes pairwise differences between the category-specific features obtained from the previous stage and then integrates complementary information across multiple feature pairs to refine node representations. Finally, we design a dual-component optimization function that synergistically combines intermediate supervision loss with the final classification objective, encouraging the network to learn robust and fine-grained node representations. Extensive experiments on multiple real-world benchmark datasets demonstrate the superior performance of CA-NodeNet over existing state-of-the-art methods. Ablation studies further validate the effectiveness of each module in contributing to overall performance gains. Full article

Person re-identification (Re-ID) has attracted considerable attention in the field of computer vision, primarily due to its critical role in video surveillance and public security applications. However, most existing Re-ID approaches rely on image-level erasing techniques, which may inadvertently remove fine-grained visual cues that are essential for accurate identification. To mitigate this limitation, we propose an effective feature erasing-based data augmentation framework that aims to explore discriminative information within individual samples and improve overall recognition performance. Specifically, we first introduce a diagonal swapping augmentation strategy to increase the diversity of the training samples. Secondly, we design a feature erasing-driven method applied to the extracted pedestrian feature to capture identity-relevant information at the feature level. Finally, extensive experiments demonstrate that our method achieves competitive performance compared to many representative approaches. Full article

To address the complex challenge of recognizing mixed handwritten text in practical scenarios such as examination papers and to overcome the limitations of existing methods that typically focus on a single category, this paper proposes MHTR, a Multimodal Handwritten Text Adaptive Recognition algorithm. The framework comprises two key components, a Handwritten Character Classification Module and a Handwritten Text Adaptive Recognition Module, which work in conjunction. The classification module performs fine-grained analysis of the input image, identifying different types of handwritten content such as Chinese characters, digits, and mathematical formula. Based on these results, the recognition module dynamically selects specialized sub-networks tailored to each category, thereby enhancing recognition accuracy. To further reduce errors caused by similar character shapes and diverse handwriting styles, a Context-aware Recognition Optimization Module is introduced. This module captures local semantic and structural information, improving the model’s understanding of character sequences and boosting recognition performance. Recognizing the limitations of existing public handwriting datasets, particularly their lack of diversity in character categories and writing styles, this study constructs a heterogeneous, integrated handwritten text dataset. The dataset combines samples from multiple sources, including Chinese characters, numerals, and mathematical symbols, and features high structural complexity and stylistic variation to better reflect real-world application needs. Experimental results show that MHTR achieves a recognition accuracy of 86.63% on the constructed dataset, significantly outperforming existing methods. Furthermore, the context-aware optimization module demonstrates strong adaptive correction capabilities in various misrecognition scenarios, confirming the effectiveness and practicality of the proposed approach for complex, multi-category handwritten text recognition tasks. Full article

Rice (Oryza sativa L.) is a staple crop for over half of the global population; however, pathogenic infections pose significant threats to its sustainable production. Although chemical pesticides are commonly employed for disease control, their prolonged usage has led to pathogen resistance, reduced effectiveness, and non-target toxicity, rendering them unsustainable for agricultural practices. Nanomaterials (NMs) present a promising alternative due to their small size, tunable release properties, and diverse mechanisms for disease resistance. This review examines how NMs can enhance rice disease management through (1) direct pathogen suppression; (2) the activation of plant defense pathways; (3) the formation of nanoscale barriers on leaves to obstruct pathogens; (4) targeted delivery and controlled release of fungicides; and (5) modulation of the microbiome to bolster resilience. Moreover, we critically analyze the agricultural potential and environmental implications of NMs, develop optimized application strategies, and, for the first time, propose the innovative ‘NMs-Rice-Soil’ Ternary System framework. This groundbreaking approach integrates nanotechnology, plant physiology, and soil ecology. The pioneering framework offers transformative solutions for sustainable crop protection, illustrating how strategically engineered NMs can synergistically enhance rice productivity, grain quality, and global food security through science-based risk management and interdisciplinary innovation. Full article

This study employs deep learning techniques, specifically BERT and Latent Dirichlet Allocation (LDA), to analyze customer satisfaction and attribute-level attention from user-generated content. By integrating these insights with Kano model surveys, we systematically rank attribute preferences and enhance decision-making accuracy. Addressing the explicit attention–implicit utility discrepancy, we extend the traditional IPA–Kano model by incorporating an attention dimension, thereby constructing a three-dimensional optimization framework with eight decision spaces. This enhanced framework enables the following: (1) fine-grained classification of customer requirements by distinguishing between an attribute’s perceived salience and its actual impact on satisfaction; (2) strategic resource allocation, differentiating between quality enhancement priorities and cognitive expectation management to maximize innovation impact under resource constraints. To validate the model, we conducted a case study on wearable watches for the elderly, analyzing 12,527 online reviews to extract 41 functional attributes. Among these, 14 were identified as improvement priorities, 9 as maintenance attributes, and 7 as low-priority features. Additionally, six cognitive management strategies were formulated to address attention–utility mismatches. Comparative validation involving domain experts and consumer interviews confirmed that the proposed IPAA–Kano model, leveraging deep learning, outperforms the traditional IPA–Kano model in classification accuracy and decision relevance. By integrating deep learning with optimization-based decision models, this research offers a practical and systematic methodology for translating customer attention and satisfaction data into actionable innovation strategies, thus providing a robust, data-driven approach to resource-efficient product development and technological innovation. Full article

Solid-state lithium batteries (SSLBs) have emerged as a promising alternative to conventional lithium-ion systems due to their superior safety profile, higher energy density, and potential compatibility with lithium metal anodes. However, a major challenge hindering their widespread deployment is the formation and growth of lithium dendrites, which compromise both performance and safety. This review provides a comprehensive and structured overview of recent advances in dendrite suppression strategies, with special emphasis on the role played by the nature of the solid electrolyte. In particular, we examine suppression mechanisms and material innovations within the three main classes of solid electrolytes: sulfide-based, oxide-based, and polymer-based systems. Each electrolyte class presents distinct advantages and challenges in relation to dendrite behavior. Sulfide electrolytes, known for their high ionic conductivity and good interfacial wettability, suffer from poor mechanical strength and chemical instability. Oxide electrolytes exhibit excellent electrochemical stability and mechanical rigidity but often face high interfacial resistance. Polymer electrolytes, while mechanically flexible and easy to process, generally have lower ionic conductivity and limited thermal stability. This review discusses how these intrinsic properties influence dendrite nucleation and propagation, including the role of interfacial stress, grain boundaries, void formation, and electrochemical heterogeneity. To mitigate dendrite formation, we explore a variety of strategies including interfacial engineering (e.g., the use of artificial interlayers, surface coatings, and chemical additives), mechanical reinforcement (e.g., incorporation of nanostructured or gradient architectures, pressure modulation, and self-healing materials), and modifications of the solid electrolyte and electrode structure. Additionally, we highlight the critical role of advanced characterization techniques—such as in situ electron microscopy, synchrotron-based X-ray diffraction, vibrational spectroscopy, and nuclear magnetic resonance (NMR)—for elucidating dendrite formation mechanisms and evaluating the effectiveness of suppression strategies in real time. By integrating recent experimental and theoretical insights across multiple disciplines, this review identifies key limitations in current approaches and outlines emerging research directions. These include the design of multifunctional interphases, hybrid electrolytes, and real-time diagnostic tools aimed at enabling the development of reliable, scalable, and dendrite-free SSLBs suitable for practical applications in next-generation energy storage. Full article

As a new type of production factor, data possesses multidimensional application value, and its pivotal role is becoming increasingly prominent in the aviation sector. Data sharing can significantly enhance the utilization efficiency of data resources and serves as one of the key tasks in building smart civil aviation. However, currently, data silos are pervasive, with vast amounts of data only being utilized and analyzed within limited scopes, leaving their full potential untapped. The challenges in data sharing primarily stem from three aspects: (1) Data owners harbor concerns regarding data security and privacy. (2) The highly dynamic and real-time nature of aviation operations imposes stringent requirements on the timeliness, stability, and reliability of data sharing, thereby constraining its scope and extent. (3) The lack of reasonable incentive mechanisms results in insufficient motivation for data owners to share. Consequently, addressing the issue of aviation big data sharing holds significant importance. Since the release of the Bitcoin whitepaper in 2008, blockchain technology has achieved continuous breakthroughs in the fields of data security and collaborative computing. Its unique characteristics—decentralization, tamper-proofing, traceability, and scalability—lay the foundation for its integration with aviation. Blockchain can deeply integrate with air traffic management (ATM) operations, effectively resolving trust, efficiency, and collaboration challenges in distributed scenarios for ATM data. To address the heterogeneous data usage requirements of different ATM stakeholders, this paper constructs a blockchain-based multi-level data security sharing architecture, enabling fine-grained management and secure collaboration. Furthermore, to meet the stringent timeliness demands of aviation operations and the storage pressure posed by massive data, this paper optimizes blockchain storage deployment and consensus mechanisms, thereby enhancing system scalability and processing efficiency. Additionally, a dual-mode data-sharing solution combining raw data sharing and model sharing is proposed, offering a novel approach to aviation big data sharing. Security and formal analyses demonstrate that the proposed solution is both secure and effective. Full article

Observing building changes in remote sensing images plays a crucial role in monitoring urban development and promoting sustainable urbanization. Mainstream change detection methods have demonstrated promising performance in identifying building changes. However, buildings have large intra-class variance and high similarity with other objects, limiting the generalization ability of models in diverse scenarios. Moreover, most existing methods only detect whether changes have occurred but ignore change types, such as new construction and demolition. To address these issues, we present a building change-type detection network (BCTDNet) based on the Segment Anything Model (SAM) to identify newly constructed and demolished buildings. We first construct a dual-feature interaction encoder that employs SAM to extract image features, which are then refined through trainable multi-scale adapters for learning architectural structures and semantic patterns. Moreover, an interactive attention module bridges SAM with a Convolutional Neural Network, enabling seamless interaction between fine-grained structural information and deep semantic features. Furthermore, we develop a change-aware attribute decoder that integrates building semantics into the change detection process via an extraction decoding network. Subsequently, an attribute-aware strategy is adopted to explicitly generate distinct maps for newly constructed and demolished buildings, thereby establishing clear temporal relationships among different change types. To evaluate BCTDNet’s performance, we construct the JINAN-MCD dataset, which covers Jinan’s urban core area over a six-year period, capturing diverse change scenarios. Moreover, we adapt the WHU-CD dataset into WHU-MCD to include multiple types of changing. Experimental results on both datasets demonstrate the superiority of BCTDNet. On JINAN-MCD, BCTDNet achieves improvements of 12.64% in IoU and 11.95% in F1 compared to suboptimal methods. Similarly, on WHU-MCD, it outperforms second-best approaches by 2.71% in IoU and 1.62% in F1. BCTDNet’s effectiveness and robustness in complex urban scenarios highlight its potential for applications in land-use analysis and urban planning. Full article