[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-4750":3,"related-4750":108},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":24,"tags":26,"primary_topic":32,"secondary_topics":38,"classification":39,"search_phrases":75,"slug":78,"view_count":22,"doi":79,"paper":80,"created_at":107},4750,"A global–local momentum contrastive framework with critical feature calibration for field plant disease identification in precision agriculture","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-74163-6","Plant diseases can substantially affect crop production and food supply stability, making accurate disease identification important for effective crop management. Although deep learning has been widely applied to plant disease recognition, model performance may degrade in complex field environments because of large intra-class variations, subtle inter-class differences, and background interference. To address these challenges, we propose GLC-Net, which integrates global–local momentum contrastive learning with critical feature calibration (CFC) for field plant disease identification. Specifically, a label-guided momentum contrastive learning mechanism is introduced to dynamically construct multiple positive pairs through a labeled queue, thereby promoting intra-class compactness and inter-class separability in the feature space. In addition, a global–local contrastive strategy based on cropped local views is designed to enhance multi-scale representation learning, enabling the model to capture both global contextual information and fine-grained pathological details. Furthermore, the proposed CFC module calibrates global representations using complementary local features, allowing the network to preserve more disease-relevant information under complex background conditions. Experiments on the Paddy and Soybean disease datasets demonstrate that GLC-Net consistently improves recognition performance across different backbone networks and compares favorably with representative existing methods. Ablation and visualization analyses further verify the contributions of the proposed components to feature discrimination and disease-region representation. These results demonstrate the effectiveness of GLC-Net for field plant disease identification, while further evaluation across additional crop species and more diverse field conditions remains necessary to establish its broader generalization capability.","植物病害会显著影响作物生产和粮食供应稳定性，因此准确的病害识别对于有效的作物管理十分重要。尽管深度学习已被广泛应用于植物病害识别，但在复杂的田间环境中，由于类内差异大、类间差异细微以及背景干扰等因素，模型性能可能会下降。为解决这些挑战，我们提出了GLC-Net，它将全局-局部动量对比学习与关键特征校准（CFC）相结合，用于田间植物病害识别。具体而言，引入了一种标签引导的动量对比学习机制，通过带标签队列动态构建多个正样本对，从而促进特征空间中的类内紧凑性和类间可分性。此外，设计了一种基于裁剪局部视图的全局-局部对比策略，以增强多尺度表征学习，使模型能够同时捕获全局上下文信息和细粒度病理细节。进一步地，所提出的CFC模块利用互补的局部特征校准全局表征，使网络能够在复杂背景条件下保留更多与病害相关的信息。在水稻和大豆病害数据集上的实验表明，GLC-Net在不同骨干网络上均能持续提升识别性能，并与代表性现有方法相比具有优势。消融实验和可视化分析进一步验证了所提出组件对特征判别和病害区域表征的贡献。这些结果证明了GLC-Net在田间植物病害识别中的有效性，但仍需在更多作物种类和更多样的田间条件下开展进一步评估，以确立其更广泛的泛化能力。",null,"Scientific Reports","2026-10-10T00:00:00Z","论文",10,false,72,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},12,21,17,14,8,1,"提出全局-局部动量对比学习与关键特征校准的GLC-Net，在水稻和大豆病害数据集上验证了田间病害识别性能提升，方法新颖但属实验室阶段成果，产业影响有限。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","农业人工智能","植物病害识别","水稻病害","大豆病害",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":40,"locked":14,"version":41,"method":42,"evidence":43},"classified","2026-10-22.2","rules",{"candidates":44,"agricultural_evidence":48},[45,53,59,67],{"code":46,"score":47,"supported":48,"title":49,"abstract":51,"tags":52},"I1",9,true,[50],"precision agriculture",[],[27],{"code":33,"score":21,"supported":48,"title":54,"abstract":55,"tags":58},[],[56,57],"深度学习","deep learning",[28,29],{"code":60,"score":61,"supported":48,"title":62,"abstract":63,"tags":66},"A1",7,[],[64,65],"水稻","paddy",[30],{"code":68,"score":69,"supported":48,"title":70,"abstract":71,"tags":74},"A4",6,[],[72,73],"大豆","soybean",[],[76,77],"GLC-Net 植物病害识别","田间植物病害 对比学习","GLC-Net植物病害识别-4750","10.1038\u002Fs41598-026-74163-6",{"doi":79,"openalex_id":81,"authors":82,"venue":10,"cited_by_count":97,"oa_url":98,"card":99,"direction":105,"ingested_from":106},"W7221328262",[83,85,87,90,92,94],{"name":84,"orcid":9},"Renyong Pan",{"name":86,"orcid":9},"Qihang Yang",{"name":88,"orcid":89},"Shaozhao Qin","https:\u002F\u002Forcid.org\u002F0009-0005-2982-4702",{"name":91,"orcid":9},"Shouwei Chen",{"name":93,"orcid":9},"Chao Yang",{"name":95,"orcid":96},"Xiaoping Wu","https:\u002F\u002Forcid.org\u002F0000-0001-6776-6973",0,"https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-74163-6_reference.pdf",{"tldr":100,"method":101,"finding":102,"direction":103,"opportunity":104},"提出GLC-Net，用全局-局部动量对比学习与关键特征校准提升田间植物病害识别。","标签引导动量对比学习、全局-局部对比策略与CFC模块，在稻田和大豆病害数据集上验","GLC-Net在不同骨干网络上稳定提升识别性能，优于代表性方法，消融与可视化验证各组件有效。","农业人工智能与决策模型","可扩展至更多作物与复杂田间条件，探索轻量化部署与跨域泛化能力。","农业遥感与作物表型","openalex","2026-10-12T23:30:24.578331Z",{"total":69,"page":22,"page_size":69,"items":109},[110,172,231,284,336,391],{"id":111,"title":112,"url":113,"summary":114,"summary_zh":115,"content":9,"source_name":116,"source_url":113,"published_at":117,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":118,"score_detail":119,"sources":122,"tags":124,"primary_topic":127,"secondary_topics":133,"classification":135,"search_phrases":152,"slug":155,"view_count":97,"doi":156,"paper":157,"created_at":171},5592,"Evaluation of Concept-Rank Collapse Audits for Computer Vision Agriculture Disease Classifiers Under Low-Label Regimes: An Interpretability Capacity Bound for Smallholder Fungicide Selectivity","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545193","Concept-based interpretability methods for plant-disease classifiers are typically evaluated by whether their concept activations correlate with human-annotated labels, but this criterion is silent on a prior question: whether the classifier's internal representation retains enough concept-rank structure for any audit to be informative. the present analysis formalises that prior question as an interpretability capacity bound. A classifier is modelled as a concept-rank operator whose singular values govern how much concept-level signal survives the projection from pixel space to class logits, and a collapse audit is defined as the ratio of retained concept-rank energy to total concept energy. Under a low-label regime, where concept annotations are scarce and the audit must be estimated from few labelled examples, the bound predicts that audit reliability degrades as a function of the concept-rank spectrum and the label budget. The analysis is a theoretical and methodological contribution: it derives the bound, states the conditions under which a collapse audit is identifiable, and characterises the operating characteristics of the audit as a decision procedure under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64). Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 52 clusters were required for 5 percentage points at 80% power. The bound reframes concept audits as capacity tests rather than fidelity tests, with consequences for how smallholder fungicide-selectivity decisions should be supported by interpretable disease classifiers.","面向植物病害分类器的基于概念的可解释性方法，通常以其概念激活是否与人工标注标签相关作为评价标准，但该标准对一个更先前的问题保持沉默：分类器的内部表征是否保留了足够的概念秩结构，从而使任何审计具有信息量。本分析将该先前问题形式化为一个可解释性容量上界。分类器被建模为概念秩算子，其奇异值支配着从像素空间到类别logits的投影中概念级信号得以保留的程度，并将坍缩审计定义为保留的概念秩能量与总概念能量之比。在低标签情形下，概念标注稀缺且审计必须由少量标注样本估计，该上界预测审计可靠性会随概念秩谱和标签预算的函数而退化。本分析是一项理论和方法论贡献：推导出该上界，陈述坍缩审计可识别的条件，并在声明的方差结构（簇间SD = 6.0，残差SD = 4.5，组内相关系数 = 0.64）下刻画审计作为决策程序的操作特征。在2000个设计单元中，针对簇数为8、16、24和32以及效应量为3、5和8个百分点的情形估计了功效；在80%功效下达到5个百分点需要52个簇。该上界将概念审计重新框定为容量检验而非保真度检验，并对可解释病害分类器应如何支持小农杀菌剂选择性决策具有启示意义。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",69,{"impact":21,"substance":18,"depth":120,"authority":17,"freshness":13,"relevant":22,"comment":121},18,"理论方法论文，提出可解释性容量上界并给出审计功效分析，对智慧农业病害识别可信度有参考价值，但属小众方法学研究、缺乏田间实证，暂不宜进入每日精选。",[123],{"name":116,"url":113},[27,28,125,29,126],"小农户","可解释性",{"code":128,"slug":129,"name":130,"path":131,"parent_code":132},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[134],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":136},{"candidates":137,"agricultural_evidence":48},[138,143,148],{"code":128,"score":17,"supported":48,"title":139,"abstract":141,"tags":142},[140],"smallholder",[140],[125],{"code":33,"score":13,"supported":48,"title":144,"abstract":146,"tags":147},[145],"computer vision",[],[28,29],{"code":46,"score":22,"supported":14,"title":149,"abstract":150,"tags":151},[],[],[27],[153,154],"农业人工智能 植物病害识别 可解释性 智慧农业","农业人工智能 植物病害识别","农业人工智能植物病害识别可解释性智慧农业-5592","10.5281\u002Fzenodo.23545193",{"doi":156,"openalex_id":158,"authors":159,"venue":116,"cited_by_count":97,"oa_url":113,"card":166,"direction":103,"ingested_from":106},"W7223137323",[160,162,164],{"name":161,"orcid":9},"Akshay Ramgoolam",{"name":163,"orcid":9},"Priya Bundhoo Appadoo",{"name":165,"orcid":9},"Jean-Noël LEspérance",{"tldr":167,"method":168,"finding":169,"direction":103,"opportunity":170},"提出可解释性容量上界，评估低标签下概念秩崩溃审计的可靠性。","将分类器建模为概念秩算子，推导容量上界并做2000格功效模拟。","审计可靠性随概念秩谱与标签预算下降；5个百分点效应需52个聚类达80%功效。","可将容量上界用于小农户杀菌剂选择性决策的可解释模型审计与标签预算设计。","2026-10-22T11:21:10.425290Z",{"id":173,"title":174,"url":175,"summary":176,"summary_zh":177,"content":9,"source_name":178,"source_url":175,"published_at":179,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":180,"score_detail":181,"sources":184,"tags":186,"primary_topic":189,"secondary_topics":190,"classification":191,"search_phrases":211,"slug":214,"view_count":97,"doi":215,"paper":216,"created_at":230},5192,"A Comparative Study of Client-Level Stabilization Techniques in Federated Learning for Robust Plant Disease Recognition","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fs26206566","The successful integration of artificial intelligence (AI) algorithms in precision agriculture provided farmers with the means to identify potential issues, such as early disease detection or pest infestations, by making crop-related predictions. The new technological developments in this field improve the functioning of traditional agriculture by reducing the dependence on specialized human expertise and facilitating greater access to diagnostic procedures. However, most studies conducted in this area rely on centralized training of AI models, which leads to increased costs and raised privacy-related concerns. Federated learning (FL) offers a compelling framework for collaboratively training machine learning (ML) models on distributed datasets without centralizing the data owned by each participant. FL offers significant advantages in the preservation of data privacy, facilitates the collaborative exploitation of geographically distributed data sources, and reduces the risks associated with centralized collection. In this paper, we investigate the current limitations and performance boundaries of existing Federated and privacy-preserving learning in distributed sensing systems for plant disease detection, in scenarios that closely resemble real-world conditions. Our analysis specifically examined how FL aggregation algorithms, such as FedAvg, FedAvgM and FedProx, work in a custom multi-client experimental setup where data is partitioned so that each participant observes only those symptom patterns associated with specific environmental and climatic conditions. We propose further partial aggregation of the local model, an approach where we send only the model backbone to the central server, while the classification head of each participant remains strictly on the client side. The construction of environmentally heterogeneity is combined with a globally representative stabilizer node designed to preserve knowledge across disease classes that may be absent from certain clients.","人工智能（AI）算法在精准农业中的成功集成，为农民提供了通过作物相关预测来识别潜在问题（如早期病害检测或虫害侵染）的手段。该领域的新技术发展通过减少对专业人类知识的依赖并促进更广泛的诊断程序获取，改善了传统农业的运作方式。然而，该领域的大多数研究依赖于AI模型的集中式训练，这导致成本增加并引发隐私相关担忧。联邦学习（FL）提供了一个引人注目的框架，可在分布式数据集上协作训练机器学习（ML）模型，而无需集中化各参与者拥有的数据。联邦学习在保护数据隐私方面具有显著优势，促进了地理上分布的数据源的协作利用，并降低了与集中收集相关的风险。本文研究了在高度接近真实条件的场景中，现有联邦学习与隐私保护学习在用于植物病害检测的分布式感知系统中的当前局限性与性能边界。我们的分析特别考察了联邦学习聚合算法（如FedAvg、FedAvgM和FedProx）在自定义多客户端实验设置中的工作方式，其中数据被划分，使得每个参与者仅观察到与特定环境和气候条件相关的症状模式。我们提出对本地模型进行进一步的部分聚合，即仅将模型骨干发送至中央服务器，而每个参与者的分类头严格保留在客户端侧。环境异质性的构建与一个全局代表性的稳定器节点相结合，该节点旨在保留某些客户端可能缺失的跨病害类别的知识。","Sensors","2026-10-16T00:00:00Z",78,{"impact":120,"substance":18,"depth":19,"authority":182,"freshness":47,"relevant":22,"comment":183},13,"该研究针对植物病害识别中的联邦学习客户端稳定化方法提出部分聚合与稳定节点方案，方法新颖、实验贴近真实场景，对智慧农业隐私计算有参考价值，值得进入每日精选。",[185],{"name":178,"url":175},[27,28,187,29,188],"联邦学习","数据隐私",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":192},{"candidates":193,"agricultural_evidence":48},[194,201,206],{"code":33,"score":17,"supported":48,"title":195,"abstract":196,"tags":200},[],[197,198,199],"机器学习","machine learning","artificial intelligence",[28,29,187],{"code":46,"score":61,"supported":48,"title":202,"abstract":203,"tags":205},[],[204,50],"精准农业",[27],{"code":207,"score":22,"supported":14,"title":208,"abstract":209,"tags":210},"I6",[],[],[188],[212,213],"联邦学习 植物病害识别","FedProx 作物病害检测","联邦学习植物病害识别-5192","10.3390\u002Fs26206566",{"doi":215,"openalex_id":217,"authors":218,"venue":178,"cited_by_count":97,"oa_url":175,"card":225,"direction":103,"ingested_from":106},"W7222586734",[219,222],{"name":220,"orcid":221},"Aurora-Felicia Cristea","https:\u002F\u002Forcid.org\u002F0009-0003-9718-065X",{"name":223,"orcid":224},"Ciprian Dobre","https:\u002F\u002Forcid.org\u002F0000-0003-4638-7725",{"tldr":226,"method":227,"finding":228,"direction":103,"opportunity":229},"比较联邦学习客户端级稳定化技术，提升植物病害识别的鲁棒性。","FedAvg\u002FFedAvgM\u002FFedProx聚合，仅上传骨干网络、分类头留本地，","部分聚合加稳定节点可缓解环境异质性与类别缺失带来的性能退化。","可探索面向非独立同分布农业数据的个性化联邦学习与通信效率优化。","2026-10-17T23:30:45.845684Z",{"id":232,"title":233,"url":234,"summary":235,"summary_zh":236,"content":9,"source_name":237,"source_url":234,"published_at":238,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":239,"score_detail":240,"sources":243,"tags":245,"primary_topic":247,"secondary_topics":248,"classification":249,"search_phrases":267,"slug":270,"view_count":22,"doi":271,"paper":272,"created_at":283},4982,"A Nightshade-SWIN Transformer-Based Framework for Enhanced Multi-Crop Plant Disease Classification","https:\u002F\u002Fdoi.org\u002F10.17148\u002Fijarcce.2026.151037","Nightshade crops (Solanaceae) face significant yield threats from phenotypically similar foliar diseases that can be difficult to distinguish during the early stages of infection.Accurate and timely disease identification is therefore essential for reducing crop losses, improving disease management, and supporting sustainable agricultural practices.While Convolutional Neural Networks (CNNs) have demonstrated strong performance in image-based plant disease classification due to their ability to extract spatial and local features, they may have limitations in capturing long-range dependencies and global contextual relationships within leaf images.These limitations become particularly important when disease symptoms exhibit subtle differences in texture, color, shape, and lesion distribution or when images are captured against complex field backgrounds.To address these challenges, we propose Nightshade-SWIN, an optimized deep learning framework based on the Swin Transformer architecture.The proposed framework utilizes a hierarchical shifted-window self-attention mechanism to effectively capture both minute lesion textures and global leaf morphology.By progressively learning multi-scale representations, Nightshade-SWIN can identify discriminative pathological characteristics while maintaining computational efficiency.The model was evaluated on a curated dataset of 5,000 multicrop images collected under variable field conditions, including differences in illumination, background complexity, leaf orientation, and disease severity.Experimental results demonstrate that Nightshade-SWIN achieves a peak classification accuracy of 94.6%, outperforming ResNet50 and EfficientNet-B0 by 4.9% and 3.4%, respectively.These results indicate the effectiveness of transformer-based feature learning for distinguishing visually similar foliar diseases.Furthermore, interpretability analysis using LIME and Grad-CAM confirms that the model primarily focuses on biologically relevant pathological regions, such as lesions, discoloration, spots, and infected leaf areas.This enhances the transparency and reliability of the proposed framework and supports its potential application in precision agriculture, automated crop monitoring, and early disease diagnosis.","茄科作物（Solanaceae）面临表型相似叶部病害的严重产量威胁，这些病害在感染早期难以区分。因此，准确及时地识别病害对于减少作物损失、改善病害管理以及支持可持续农业实践至关重要。尽管卷积神经网络（CNN）因其提取空间和局部特征的能力而在基于图像的植物病害分类中表现出色，但它们在捕捉叶片图像中的长距离依赖关系和全局上下文关联方面可能存在局限。当病害症状在纹理、颜色、形状和病斑分布上表现出细微差异，或图像在复杂田间背景下拍摄时，这些局限尤为突出。为应对这些挑战，我们提出了Nightshade-SWIN，一种基于Swin Transformer架构的优化深度学习框架。该框架利用分层移位窗口自注意力机制，有效捕捉微小的病斑纹理和全局叶片形态。通过逐步学习多尺度表征，Nightshade-SWIN能够在保持计算效率的同时识别具有判别力的病理特征。该模型在包含5,000张多作物图像的数据集上进行了评估，这些图像采集自多变田间条件，包括光照、背景复杂度、叶片朝向和病害严重程度的差异。实验结果表明，Nightshade-SWIN达到了94.6%的最高分类准确率，分别优于ResNet50和EfficientNet-B0达4.9%和3.4%。这些结果表明了基于Transformer的特征学习在区分视觉相似叶部病害方面的有效性。此外，使用LIME和Grad-CAM进行的可解释性分析证实，该模型主要关注生物学相关的病理区域，如病斑、变色、斑点及受感染叶片区域。这增强了所提框架的透明度和可靠性，并支持其在精准农业、自动化作物监测和早期病害诊断中的潜在应用。","IJARCCE","2026-10-14T00:00:00Z",68,{"impact":17,"substance":241,"depth":19,"authority":13,"freshness":47,"relevant":22,"comment":242},20,"该论文提出基于Swin Transformer的茄科作物病害分类框架，在5000张多作物图像上达到94.6%准确率，方法新颖、数据规模可观，对智慧农业病害识别有参考价值，但属细分领域学术进展，公共影响有限。",[244],{"name":237,"url":234},[27,28,204,29,246],"茄科作物",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":250},{"candidates":251,"agricultural_evidence":48},[252,256,260],{"code":46,"score":21,"supported":48,"title":253,"abstract":254,"tags":255},[],[204,50],[27,204],{"code":33,"score":21,"supported":48,"title":257,"abstract":258,"tags":259},[],[56,57],[28,29],{"code":261,"score":262,"supported":48,"title":263,"abstract":264,"tags":266},"H8",3,[],[265],"可持续农业",[],[268,269],"Nightshade-SWIN 病害分类","Swin Transformer 植物病害","Nightshade-SWIN病害分类-4982","10.17148\u002Fijarcce.2026.151037",{"doi":271,"openalex_id":273,"authors":274,"venue":237,"cited_by_count":97,"oa_url":234,"card":277,"direction":282,"ingested_from":106},"W7221746668",[275],{"name":276,"orcid":9},"Dr. Barkha Joshi",{"tldr":278,"method":279,"finding":280,"direction":103,"opportunity":281},"提出Nightshade-SWIN框架，用Swin Transformer提升茄科多作物叶片病害分类","基于Swin Transformer的移位窗口自注意力，5000张田间多作物图像","分类准确率达94.6%，优于ResNet50和EfficientNet-B0，且模型关注病斑等生物相","可探索轻量化Transformer在移动端实时病害诊断及跨作物迁移学习，结合多模态数据提升田间泛化能力。","智慧农业 \u002F 农业物联网","2026-10-15T23:30:08.177373Z",{"id":285,"title":286,"url":287,"summary":288,"summary_zh":289,"content":9,"source_name":290,"source_url":287,"published_at":291,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":292,"score_detail":293,"sources":296,"tags":298,"primary_topic":301,"secondary_topics":302,"classification":303,"search_phrases":319,"slug":322,"view_count":262,"doi":323,"paper":324,"created_at":335},4907,"PEAS-DeiT: Multi-Scale Spatial Attention Enhanced DeiT for Pea Leaf Disease Classification","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fapp162010113","Early and accurate detection of pea diseases is essential for reducing crop losses and supporting precision agriculture. Although deep learning methods have shown good performance in plant disease classification, CNN-based models may have limitations in learning long-range relationships and detailed disease features. In this study, an explainable transformer-based framework, named PEAS-DeiT, is proposed for automated pea leaf disease classification. The proposed approach integrates a multi-scale spatial attention mechanism into the Data-efficient Image Transformer (DeiT) architecture to enhance the representation of disease-related visual features. A unified pea leaf disease dataset (UPLDD) was constructed by combining multiple publicly available datasets, covering six classes, including five disease categories and healthy samples. The proposed model was evaluated against CNN-based and transformer-based architectures using five-fold cross-validation. Furthermore, explainable artificial intelligence (XAI) analysis was performed to investigate the regions contributing to model predictions. Experimental results demonstrate that PEAS-DeiT achieved an accuracy of 93.11%, showing competitive performance compared with the evaluated CNN and transformer-based models. The gradient-based saliency analysis also revealed localized response patterns in several evaluated samples, providing qualitative insight into the image regions contributing to model predictions. Statistical analyses were conducted to evaluate the performance differences among the compared models. These findings indicate that PEAS-DeiT provides a competitive approach for pea leaf disease classification within the evaluated experimental setting.","豌豆病害的早期准确检测对于减少作物损失和支持精准农业至关重要。尽管深度学习方法在植物病害分类中表现出良好性能，但基于卷积神经网络（CNN）的模型在学习长程关系和病害细节特征方面可能存在局限。本研究提出了一种基于Transformer的可解释框架PEAS-DeiT，用于豌豆叶片病害的自动化分类。该方法将多尺度空间注意力机制集成到数据高效图像Transformer（DeiT）架构中，以增强病害相关视觉特征的表征能力。通过整合多个公开数据集，构建了统一的豌豆叶片病害数据集（UPLDD），涵盖六个类别，包括五类病害和健康样本。采用五折交叉验证，将所提模型与基于CNN和基于Transformer的架构进行了评估比较。此外，进行了可解释人工智能（XAI）分析，以探究对模型预测有贡献的区域。实验结果表明，PEAS-DeiT达到了93.11%的准确率，与所评估的基于CNN和基于Transformer的模型相比展现出有竞争力的性能。基于梯度的显著性分析还揭示了若干评估样本中的局部响应模式，为理解对模型预测有贡献的图像区域提供了定性见解。通过统计分析评估了所比较模型之间的性能差异。这些发现表明，在所评估的实验设置下，PEAS-DeiT为豌豆叶片病害分类提供了一种有竞争力的方法。","Applied Sciences","2026-10-13T00:00:00Z",64,{"impact":21,"substance":120,"depth":294,"authority":182,"freshness":47,"relevant":22,"comment":295},16,"提出多尺度空间注意力增强的DeiT模型用于豌豆叶病害分类，构建六类统一数据集并达93.11%准确率，方法有创新但属实验室验证阶段，产业影响有限。",[297],{"name":290,"url":287},[27,28,299,29,300],"可解释AI","豌豆",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":304},{"candidates":305,"agricultural_evidence":48},[306,310,314],{"code":33,"score":17,"supported":48,"title":307,"abstract":308,"tags":309},[],[56,57,199],[28,299,29],{"code":46,"score":61,"supported":48,"title":311,"abstract":312,"tags":313},[],[204,50],[27],{"code":315,"score":22,"supported":14,"title":316,"abstract":317,"tags":318},"A5",[],[],[300],[320,321],"PEAS-DeiT 豌豆叶病害","DeiT 植物病害分类","PEAS-DeiT豌豆叶病害-4907","10.3390\u002Fapp162010113",{"doi":323,"openalex_id":325,"authors":326,"venue":290,"cited_by_count":97,"oa_url":287,"card":330,"direction":103,"ingested_from":106},"W7221612639",[327],{"name":328,"orcid":329},"Fatma Ozge Ozkok","https:\u002F\u002Forcid.org\u002F0000-0002-1421-4670",{"tldr":331,"method":332,"finding":333,"direction":103,"opportunity":334},"提出PEAS-DeiT模型，将多尺度空间注意力融入DeiT，实现豌豆叶片病害分类。","多尺度空间注意力增强DeiT，构建UPLDD数据集，五折交叉验证与XAI分析。","PEAS-DeiT达到93.11%准确率，优于对比的CNN和Transformer模型。","可探索轻量化Transformer在移动端病害识别中的部署，并扩展多作物跨域泛化能力。","2026-10-14T23:30:46.091813Z",{"id":337,"title":338,"url":339,"summary":340,"summary_zh":341,"content":9,"source_name":342,"source_url":339,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":343,"score_detail":344,"sources":347,"tags":349,"primary_topic":351,"secondary_topics":352,"classification":353,"search_phrases":364,"slug":367,"view_count":22,"doi":368,"paper":369,"created_at":390},4743,"Image Processing-Based Efficient Hybrid Feature Extraction Approach for Plant Leaf Disease Recognition","https:\u002F\u002Fdoi.org\u002F10.14445\u002F23488549\u002Fijece-v13i9p107","In this paper, a novel hybrid method for plant leaf disease recognition is presented, combining image processing techniques with Deep Learning and handcrafted feature analysis. Leaf images are processed using a Convolutional Neural Network (CNN) to automatically extract hierarchical features, while handcrafted features based on leaf shape, color, and texture provide complementary disease-related information. The combined feature set is fed into a multi-class classifier to accurately identify specific disease types. Evaluation on a diverse dataset of healthy and diseased leaves demonstrates an average accuracy of 98%, outperforming approaches that rely solely on deep learning or handcrafted features. The results highlight the effectiveness of integrating deep and handcrafted features, offering a reliable, scalable solution for precision agriculture and serving as a foundation for future research in automated plant disease detection.","本文提出了一种新颖的植物叶片病害识别混合方法，将图像处理技术与深度学习及手工特征分析相结合。利用卷积神经网络（CNN）对叶片图像进行处理，以自动提取层次化特征，同时基于叶片形状、颜色和纹理的手工特征提供了与病害相关的补充信息。将组合后的特征集输入多类分类器，以准确识别具体的病害类型。在包含健康叶片和患病叶片的多样化数据集上进行评估，结果表明平均准确率达到98%，优于仅依赖深度学习或手工特征的方法。结果凸显了融合深度特征与手工特征的有效性，为精准农业提供了一种可靠且可扩展的解决方案，并为自动化植物病害检测的未来研究奠定了基础。","International Journal of Electronics and Communication Engineering",63,{"impact":17,"substance":120,"depth":345,"authority":13,"freshness":21,"relevant":22,"comment":346},15,"提出深度学习与手工特征融合的叶片病害识别方法，准确率达98%，对精准农业有参考价值，但属实验室论文，产业影响有限。",[348],{"name":342,"url":339},[27,28,29,350],"图像识别",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":354},{"candidates":355,"agricultural_evidence":48},[356,360],{"code":33,"score":47,"supported":48,"title":357,"abstract":358,"tags":359},[],[56,57],[28,350,29],{"code":46,"score":61,"supported":48,"title":361,"abstract":362,"tags":363},[],[204,50],[27],[365,366],"植物叶片病害识别 图像处理","CNN 特征提取 病害检测","植物叶片病害识别图像处理-4743","10.14445\u002F23488549\u002Fijece-v13i9p107",{"doi":368,"openalex_id":370,"authors":371,"venue":342,"cited_by_count":97,"oa_url":384,"card":385,"direction":282,"ingested_from":106},"W7221322423",[372,374,377,380,382],{"name":373,"orcid":9},"Arundhati Bora",{"name":375,"orcid":376},"Parismita Sarma","https:\u002F\u002Forcid.org\u002F0000-0003-1904-820X",{"name":378,"orcid":379},"Dankan Gowda V","https:\u002F\u002Forcid.org\u002F0000-0003-0724-0333",{"name":381,"orcid":9},"Manash Pratim Bhuyan",{"name":383,"orcid":9},"Jumi Sarmah","https:\u002F\u002Fwww.internationaljournalssrg.org\u002F..\u002FIJECE\u002F2026\u002FVolume13-Issue9\u002FIJECE-V13I9P107.pdf",{"tldr":386,"method":387,"finding":388,"direction":103,"opportunity":389},"提出融合CNN深度特征与手工特征的方法，用于植物叶片病害识别。","CNN自动提取特征，结合叶片形状、颜色、纹理手工特征，多类分类器。","平均准确率达98%，优于仅用深度学习或手工特征的方法。","可探索轻量化混合特征模型在移动端或边缘设备的实时病害识别应用。","2026-10-12T23:30:07.942335Z",{"id":392,"title":393,"url":394,"summary":395,"summary_zh":396,"content":9,"source_name":10,"source_url":394,"published_at":397,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":398,"score_detail":399,"sources":401,"tags":403,"primary_topic":405,"secondary_topics":406,"classification":407,"search_phrases":419,"slug":422,"view_count":22,"doi":423,"paper":424,"created_at":448},4425,"Fuzzy-guided hyperparameter optimization for efficient CNN-based plant disease classification","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-64880-3","Abstract Deep Convolutional Neural Networks (DCNNs) have exhibited outstanding performance in plant disease classification. Nevertheless, their usefulness is very susceptible to hyperparameter selection. Efficient Hyper-Parameter Optimization (HPO) remains a fundamental challenge in deploying deep learning models for plant disease classification, as implementing extensive search techniques such as grid search, Bayesian Optimization (BO), and evolutionary algorithms requires considerable computational cost. Such conventional HPO approaches frequently rely on single-objective fitness criteria and demand substantial computational resources. Moreover, these approaches often lack interpretability in decision making. In this study, we propose an HPO based on a Fuzzy Inference System (FIS) approach that integrates a Mamdani FIS into the hyperparameter selection process of a DCNN. It maintains classification accuracy, while speeding up the hyperparameter selection process. The candidate configurations are assessed based on several factors, including validation accuracy, validation loss, and training time, rather than being assessed based on just one factor. These factors are combined into a single, crisp fitness score using fuzzy linguistic rules, providing transparent, interpretable, and multi-objective decision making. In order to minimize computational cost, candidate hyperparameters are first assessed using a quick partial training method that makes use of frozen convolutional layers and warm-started weights. A Mamdani FIS is then used to aggregate performance metrics including accuracy, loss, and training time, producing a crisp score for candidate ranking. After that, training of the chosen hyperparameters is exposed to fully end-to-end training, resulting in superior classification performance with respect to the state-of-the-art exhaustive search techniques. Experiments have been executed on four public plant disease datasets, using an AlexNet-based architecture, where the first experiments have been performed on a baseline model with a randomly-chosen hyperparameter configuration before the fuzzy-guided optimization process is initiated. Then, the model is performed with the best optimized hyperparameter configuration selected by the fuzzy-guided optimization process, demonstrating up to 29% improvement in accuracy, precision, recall and F1-score, along with reliable and comprehensible hyperparameter ranking and optimization. The achieved results show that while keeping an acceptable computing cost, the introduced fuzzy-guided technique consistently outperforms baseline model configurations with respect to all evaluation metrics, yielding greater classification accuracy and enhanced generalization. Hence, our suggested approach provides a lightweight, comprehensible, and efficient solution for hyperparameter tuning in the plant disease classification domain when compared to conventional HPO techniques, offering a practical direction for resource-efficient model optimization in precision agriculture that enables quick deployment of deep learning systems.","深度卷积神经网络（DCNN）在植物病害分类中表现出卓越的性能。然而，其有效性极易受到超参数选择的影响。高效超参数优化（HPO）仍然是将深度学习模型部署于植物病害分类中的一个基本挑战，因为实施网格搜索、贝叶斯优化（BO）和进化算法等大规模搜索技术需要相当大的计算成本。这类传统HPO方法通常依赖单目标适应度标准，并且需要大量计算资源。此外，这些方法在决策过程中往往缺乏可解释性。在本研究中，我们提出了一种基于模糊推理系统（FIS）的HPO方法，将Mamdani FIS集成到DCNN的超参数选择过程中。该方法在保持分类精度的同时，加快了超参数选择过程。候选配置基于多个因素进行评估，包括验证准确率、验证损失和训练时间，而不是仅基于单一因素进行评估。这些因素通过模糊语言规则组合为一个单一的清晰适应度分数，从而提供透明、可解释且多目标的决策。为了最小化计算成本，候选超参数首先通过一种快速部分训练方法进行评估，该方法利用冻结的卷积层和热启动权重。随后，使用Mamdani FIS聚合包括准确率、损失和训练时间在内的性能指标，生成用于候选排序的清晰分数。之后，对所选超参数进行完全端到端训练，从而相对于最先进的穷举搜索技术取得更优的分类性能。实验在四个公开植物病害数据集上执行，采用基于AlexNet的架构，其中在启动模糊引导优化过程之前，首先在随机选择超参数配置的基线模型上进行实验。然后，使用由模糊引导优化过程选出的最佳优化超参数配置运行模型，在准确率、精确率、召回率和F1分数上最高提升29%，同时实现可靠且可理解的超参数排序与优化。所获得的结果表明","2026-10-06T00:00:00Z",74,{"impact":345,"substance":241,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":400},"提出模糊推理系统驱动的多目标超参数优化方法，在四个公开植物病害数据集上显著提升CNN分类性能，兼具方法新颖性与农业落地价值。",[402],{"name":10,"url":394},[27,28,204,29,404],"超参数优化",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":408},{"candidates":409,"agricultural_evidence":48},[410,414],{"code":33,"score":21,"supported":48,"title":411,"abstract":412,"tags":413},[],[56,57],[28,29],{"code":46,"score":415,"supported":48,"title":416,"abstract":417,"tags":418},5,[],[50],[27,204],[420,421],"CNN 植物病害分类","模糊推理 超参数优化","CNN植物病害分类-4425","10.1038\u002Fs41598-026-64880-3",{"doi":423,"openalex_id":425,"authors":426,"venue":10,"cited_by_count":97,"oa_url":442,"card":443,"direction":282,"ingested_from":106},"W7220695608",[427,429,431,434,437,439],{"name":428,"orcid":9},"Medhat Hamdy",{"name":430,"orcid":9},"Heba M. Elhoseny",{"name":432,"orcid":433},"Fathi E. Abd El‐Samie","https:\u002F\u002Forcid.org\u002F0000-0001-8749-9518",{"name":435,"orcid":436},"Adel S. El‐Fishawy","https:\u002F\u002Forcid.org\u002F0000-0003-1567-457X",{"name":438,"orcid":9},"Sami A. El‐Dolil",{"name":440,"orcid":441},"Elhossiny Ibrahim","https:\u002F\u002Forcid.org\u002F0000-0001-5421-2170","https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-64880-3_reference.pdf",{"tldr":444,"method":445,"finding":446,"direction":103,"opportunity":447},"提出模糊推理系统引导的CNN超参数优化方法，用于植物病害分类，兼顾精度与效率。","Mamdani模糊推理系统结合部分训练与热启动权重，在四个公开病害数据集上优化A","相比传统穷举搜索，精度等指标最高提升29%，且超参数排序可解释、计算成本更低。","可将模糊多目标HPO迁移到轻量边缘设备或跨作物迁移学习场景，探索动态超参数自适应机制。","2026-10-08T23:30:12.271848Z"]