Applications in Plant Sciences最新文献

{"title":"A comprehensive review of detection, optical character recognition, and structuring methods for herbarium label digitization.","authors":"Hiba Abbad, Hanane Ariouat, Nicolas Turenne, Eric Chenin, Abderrazak Sebaa, Edi Prifti, Jean-Daniel Zucker, Youcef Sklab","doi":"10.1002/aps3.70084","DOIUrl":"https://doi.org/10.1002/aps3.70084","url":null,"abstract":"<p><p>Global herbarium digitization has produced over 50 million specimen images, many of which contain handwritten or printed labels with rich metadata such as species names, collection dates, localities, and ecological notes. Manual transcription of this information is labor intensive, error prone, and not scalable, which has motivated the development of automated solutions. This systematic review examines studies published between 2018 and 2026, identified through electronic database searches and backward reference screening, focusing on deep learning and computer vision approaches for herbarium label digitization. The reviewed methods include convolutional neural networks, transformer-based models, large language models, and open-source pipelines for optical character recognition and structured metadata extraction from historical and handwritten botanical records. The results highlight significant advances in label detection, text recognition accuracy, and metadata structuring enabled by deep learning techniques, although reported performance varies considerably depending on dataset quality, handwriting complexity, and domain adaptation strategies. Despite this progress, substantial challenges remain in handling historical handwriting styles, heterogeneous label layouts, and multilingual collections. Future research should emphasize the creation of curated benchmark datasets and the development of scalable, robust artificial intelligence (AI) pipelines to enhance the reliability, efficiency, and generalizability of automated herbarium digitization systems.</p>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":" ","pages":"e70084"},"PeriodicalIF":3.5,"publicationDate":"2026-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13742598/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149369848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Characteristics of chlorophyll fluorescence in the light-adapted state as indicators of plant leaf relative water content.","authors":"Junqing Chen, Lijiang Fu, Ya Guo, Felix B Fritschi, Jinglu Tan","doi":"10.1002/aps3.70090","DOIUrl":"https://doi.org/10.1002/aps3.70090","url":null,"abstract":"<p><strong>Premise: </strong>Chlorophyll fluorescence (ChlF) in the light-adapted state is a convenient indicator of plant water status but is heavily reliant on nonphotochemical quenching (NPQ). The usefulness of ChlF as an indicator can be enhanced by reducing NPQ effects.</p><p><strong>Methods: </strong>ChlF and relative water content (RWC) were measured from leaves of <i>Arabidopsis thaliana</i> and <i>Spinacia oleracea</i> while the leaf samples progressively lost water. ChlF characteristics were then reformulated and tested as indicators of RWC via regression analysis.</p><p><strong>Results: </strong>We demonstrate that measured NPQ varies and is by itself not a consistent indicator of RWC of plant leaves. Using NPQ as a scaler in ChlF-based photochemical efficiency measures resulted in two NPQ-adjusted ChlF characteristics as indicators of RWC: photosystem II photochemical quenching (PSII PQ) and an NPQ-scaled PSII photochemical efficiency (Y(II)<i>'</i>) measure. The results were experimentally demonstrated in <i>A. thaliana</i>, then validated in <i>A. thaliana</i> with varied illumination intensities, and finally validated in another plant (<i>S. oleracea</i>). Further analysis revealed a linear correlation between steady-state ChlF (<i>F</i> _<i>s</i> ) and maximum fluorescence (<i>F</i> _<i>m</i> <i>'</i>) across different RWCs and a time constant to account for the NPQ effect.</p><p><strong>Discussion: </strong>This research shows that ChlF characteristics are affected by NPQ during water loss and that adjusting PSII efficiency measures based on NPQ provides enhanced effectiveness in indicating plant water status.</p>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":" ","pages":"e70090"},"PeriodicalIF":3.5,"publicationDate":"2026-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13740741/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149369849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PhenoFECT: A Python package for forecasting and exploring plant phenology under climate change.","authors":"Songwon Kim, In-Geol Choi","doi":"10.1002/aps3.70088","DOIUrl":"https://doi.org/10.1002/aps3.70088","url":null,"abstract":"<p><strong>Premise: </strong>Climate change and global warming significantly affect the timing of plant flowering in spring. Accurately predicting these phenological events is crucial for understanding plant biology and agriculture. The chill-day model (CDM) based on seasonal temperature changes has been a reliable model for predicting phenological activities; however, recent global warming has challenged its accuracy due to extreme weather variability, necessitating the adjustment of the CDM parameters to reflect recent climate change. We developed an open-source Python package, PhenoFECT (Phenology Forecasting Exploring ClimaTe), that refines temperature-based phenology models. The package enables efficient phenological analysis for open and user-specified data.</p><p><strong>Methods and results: </strong>The Python package embeds open meteorological data collected over 53 years (1973-2025) in Korea for cherry (<i>Prunus yedoensis</i>), golden-bell (<i>Forsythia koreana</i>), and azalea (<i>Rhododendron mucronulatum</i>). PhenoFECT was utilized to adjust the parameters of the CDM to reflect recent climate changes. The refined CDM improved the prediction accuracy of flowering at the average root mean square error (RMSE) of 5.74 days for three angiosperms.</p><p><strong>Conclusions: </strong>PhenoFECT is a versatile tool for accurate forecasting, exploration, and visualization across diverse open meteorological data. Exploring data using PhenoFECT revealed that the CDM can be influenced by geographic features and global climate phenomena.</p>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":" ","pages":"e70088"},"PeriodicalIF":3.5,"publicationDate":"2026-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13722615/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149317363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multi-task deep learning-based approach to identify and assess dry root rot disease in chickpea.","authors":"Piyush Priya, Joe Johnson, Athimoolam Durgadevi, Navaneeth Vinayachandran, Meghna Pillai, Bharat Maitraya, Muthukumar V Bagavathiannan, Muthappa Senthil-Kumar","doi":"10.1002/aps3.70075","DOIUrl":"https://doi.org/10.1002/aps3.70075","url":null,"abstract":"<p><strong>Premise: </strong>Dry root rot (DRR), an emerging global disease exacerbated by drought, affects a wide range of crops. The disease symptoms are manually scored for large-scale germplasm screening to identify disease-resistant varieties and test transgenic lines and other varieties. Despite its use in the past, manual disease identification is not only cumbersome but also often incapable of distinguishing between similar stress phenotypes. Manual disease scoring is tedious, time-consuming, and often error-prone, especially when screening large numbers of plants. To overcome this limitation, we developed a machine learning model that employs a learning-based approach using convolutional neural networks for automated feature extraction and image-based identification and severity assessment of DRR.</p><p><strong>Methods: </strong>Our models were trained, tested, and validated using an in-house dataset of 10,800 unique RGB images. A total of 51 deep learning models, including single-task and multi-task models, were developed using the Xception and Vision Transformer (ViT) architectures with standard image augmentations, and then validated on stratified independent test sets to demonstrate robustness across camera, scanner, and microscope modalities.</p><p><strong>Results: </strong>The best multi-task ViT model, capable of scoring across combined datasets, achieved 94% accuracy in distinguishing DRR from the control and had a root mean square error of 1 in predicting DRR severity score.</p><p><strong>Discussion: </strong>This model's metrics, including confusion matrix performance, underscore its potential as an efficient and accessible tool for large-scale germplasm screening for DRR, overcoming the challenges of manual scoring.</p>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":" ","pages":"e70075"},"PeriodicalIF":3.5,"publicationDate":"2026-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13722823/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149317362","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction to “The challenges of growing orchids from seeds for conservation: An assessment of asymbiotic techniques”","authors":"","doi":"10.1002/aps3.70079","DOIUrl":"10.1002/aps3.70079","url":null,"abstract":"<p>Jolman, D., M. I. Batalla, A. Hungerford, P. Norwood, N. Tait, and L. E. Wallace. 2022. The challenges of growing orchids from seeds for conservation: An assessment of asymbiotic techniques. <i>Applications in Plant Sciences</i> 10(5): e11496. https://doi.org/10.1002/aps3.11496</p><p>In the original publication of this article, an error occurred in the formatting of the name of the author Dr. Jaime A. Teixeira da Silva in a cited reference.</p><p>The corrected reference should read:</p><p>Fu, Y. Y., N. Jiang, K. L. Wu, J. X. Zhang, J. A. Teixeira da Silva, J. Duan, H. T. Liu, and S. J. Zeng. 2016. Stimulatory effects of sodium hypochlorite and ultrasonic treatments on tetrazolium staining and seed germination in vitro of <i>Paphiopedilum</i> SCBG Red Jewel. <i>Seed Science and Technology</i> 44: 77–90. https://doi.org/10.15258/sst.2016.44.1.13</p><p>We apologize for this error.</p>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13507636/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148824365","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PbAT: A user-friendly R/Shiny platform for data-driven decision support in crop improvement","authors":"Krishnan P. Abhijith,&nbsp;K. K. Vinod,&nbsp;Ranjith Kumar Ellur,&nbsp;K. T. Ravikiran,&nbsp;Rachit Kumar Saxena,&nbsp;Vignesh Muthusamy,&nbsp;S. Gopala Krishnan","doi":"10.1002/aps3.70068","DOIUrl":"10.1002/aps3.70068","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Premise</h3>\u0000 \u0000 <p>Modern plant breeding requires robust analysis of complex, multi-environment datasets to identify superior genotypes. Although R offers powerful statistical packages, their use often requires programming expertise. Existing graphical user interface (GUI)-based tools are either limited in scope or not openly accessible, leaving breeders without a comprehensive solution.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods and Results</h3>\u0000 \u0000 <p>We introduce PbAT (Plant breeding Analytical Tools), an open-access Shiny application that unifies multiple R packages into a seamless, code-free workflow. PbAT enables trial design, data curation, experimental design analysis, stability testing, multivariate methods, and mating design analysis. Outputs include variance components, best linear unbiased estimators (BLUEs), best linear unbiased predictors (BLUPs), ANOVA tables, and publication-ready plots, accompanied by automated plain-language interpretations and model equations. Its modular architecture supports future integration of additional tools. PbAT is available online as a web application (https://pbat.online) and as an R package (https://github.com/abhijithkpgen/PbAT).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>By integrating powerful statistical methods with intuitive interfaces, PbAT streamlines plant breeding analytics, enhances reproducibility, and accelerates data-driven decision-making in crop improvement.</p>\u0000 </section>\u0000 </div>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13720341/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149294199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-time monitoring of root dielectric properties for assessing crop plant damage caused by foliar application of glyphosate","authors":"Imre Cseresnyés,&nbsp;Anna Füzy,&nbsp;Tünde Takács","doi":"10.1002/aps3.70057","DOIUrl":"https://doi.org/10.1002/aps3.70057","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Premise</h3>\u0000 \u0000 <p>There is a knowledge gap regarding how foliar injury and restricted water uptake can be detected by measuring root dielectric response. This pot study nondestructively evaluated the efficiency of real-time dielectric measurement to monitor the effects of glyphosate spraying.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Root dielectric properties were recorded on a minute scale in control and glyphosate-treated maize, cucumber, and pea. Chlorophyll, stomatal conductance, and biomass measurements were taken to interpret the dielectric changes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Electrical capacitance and conductance varied diurnally due to the circadian regulation of water uptake and hydraulic conductance. Glyphosate application reduced capacitance, indicating the impeded root growth and activity caused by impaired amino acid synthesis, foliar damage, and restricted transpiration. The dissipation factor decreased in response to glyphosate due to impeded apoplastic water flow, suppressed root lignification, and hampered water absorption. The enhanced leaf and root hydraulic resistance caused by glyphosate was manifested in sharply reduced electrical conductance. Changes in the species’ dielectric response were consistent with physiological symptoms and biomass loss.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Discussion</h3>\u0000 \u0000 <p>Real-time dielectric measurement proved suitable for the nondestructive monitoring of plant responses to foliar stress through altered root traits. This method could be employed to evaluate herbicide tolerance in crops and to develop and determine dosage of herbicide ingredients.</p>\u0000 </section>\u0000 </div>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/aps3.70057","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149311503","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Turning a new leaf: PhenoVision provides leaf phenology data at the global scale","authors":"Erin L. Grady,&nbsp;Ellen G. Denny,&nbsp;Carrie E. Seltzer,&nbsp;John Deck,&nbsp;Daijiang Li,&nbsp;Russell Dinnage,&nbsp;Robert P. Guralnick","doi":"10.1002/aps3.70067","DOIUrl":"10.1002/aps3.70067","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Premise</h3>\u0000 \u0000 <p>Plant phenology dictates many aspects of community function and ecosystem dynamics. Yet, global phenology data are still limited, especially in areas lacking monitoring programs. Here we present a new data resource, PhenoVision–Leaf, which extends a computer vision pipeline utilizing iNaturalist digital image vouchers to produce global-scale leaf phenophase data for deciduous woody genera.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We first discuss our implementation of a new human annotation framework for leaf phenology on iNaturalist, aligning with phenophase definitions used by the larger phenology community. We then showcase the use of 165,988 crowdsourced annotated records to train a Vision Transformer model with a two-stage regime to maximize precision across single- and multi-image records. This approach extends PhenoVision from scoring individual images to aggregating at the iNaturalist record level, better aligning with human annotation processes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Post-hoc validation showed high performance for detecting present green and colored leaves (&gt;98% precision) and reasonable precision for breaking leaf buds (&gt;87% precision). Applying PhenoVision–Leaf to over six million iNaturalist records yielded 5.6 million record-level phenology observations across 6500 species and 57 families, filling geographic and taxonomic gaps.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Discussion</h3>\u0000 \u0000 <p>These data, now accessible through the Phenobase web application, establish a foundation for near real-time monitoring of leaf phenology, supporting global-scale synthesis analyses.</p>\u0000 </section>\u0000 </div>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13720346/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149294192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mapping Lupinus polyphyllus density and distribution along road verges using unmanned aerial vehicle (UAV)–based remote sensing","authors":"Elin L. Blomqvist,&nbsp;Jan-Olov Andersson,&nbsp;R. Lutz Eckstein,&nbsp;Jan Haas","doi":"10.1002/aps3.70080","DOIUrl":"10.1002/aps3.70080","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Premise</h3>\u0000 \u0000 <p>Road verges function as refuges for semi-natural species, but they can also facilitate the spread of non-native plants such as <i>Lupinus polyphyllus</i>. Unmanned aerial vehicle (UAV)–based remote sensing is a promising tool for mapping these species; however, its application in roadside contexts remains limited.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We developed and evaluated a UAV-based workflow to detect and map lupine presence and density along roads. Red-green-blue (RGB) imagery was processed into orthomosaics and classified using object-based and pixel-based approaches with support vector machines and random forest algorithms under multiple classification schemes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The highest accuracy (88.8%) was achieved using pixel-based random forest classification and a simplified scheme. Flowering purple lupine showed strong classification performance (F1 score = 96.3%), while early-stage greenish individuals were harder to detect (74.5%). Spatial heatmap comparison revealed low overestimation (1.79%) and underestimation (1.47%).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Discussion</h3>\u0000 \u0000 <p>Assuming optimal phenological conditions, UAV-based mapping provides a scalable and cost-efficient complement to ground surveys for the management of <i>L. polyphyllus</i> within infrastructure areas.</p>\u0000 </section>\u0000 </div>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13559118/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148928833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating machine learning, deep learning, and image analysis for seed species classification","authors":"Jonathan Ashworth,&nbsp;Eloise Barrett,&nbsp;Rebonto Haque,&nbsp;Ben Wildey,&nbsp;Anthony Hall,&nbsp;Felix Shaw,&nbsp;Jose J. De Vega","doi":"10.1002/aps3.70072","DOIUrl":"https://doi.org/10.1002/aps3.70072","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Premise</h3>\u0000 \u0000 <p>The growing demand for wildflower seeds in ecological restoration requires reliable species identification, yet current market products often contain heterogeneous species. As seed identification is labor-intensive and requires advanced botanical knowledge, we evaluated multiple segmentation and classification approaches to determine which combination performed better for identifying species and quantifying their frequencies under real-life constraints.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We generated seed images from common wildflower species using a flatbed scanner. Artificial intelligence (AI) and non-AI segmentation pipelines were compared. We evaluated over 20 classifiers on tabulated features and convolutional neural networks (CNNs) on single seed image inputs. Open-set conditions were simulated using “mock seed mixes” with previously unseen taxa using decision thresholds.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>With tabulated data, the XGBoost and Multi-layer Perceptron (MLP) models exceeded F1 &gt; 0.96, while AutoGluon's CNN and ensemble models reached F1 &gt; 0.97. ResNet-50 achieved an F1 score over 0.99 on single-seed images obtained with Cellpose segmentation. Random forest achieved the highest accuracy (0.92) with unseen species in open-set classification.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Discussion</h3>\u0000 \u0000 <p>Deep learning segmentation substantially enhanced CNN accuracy, yet tabulated-feature machine learning models remained competitive and efficient. Threshold-based rejection enabled robust open-set classification, with random forest outperforming deeper models in rejecting unseen seeds. No single method was universally optimal; the best strategy depends on whether the task involves closed-set or open-set classification and on trade-offs between accuracy and computational cost.</p>\u0000 </section>\u0000 </div>","PeriodicalId":8022,"journal":{"name":"Applications in Plant Sciences","volume":"14 5","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/aps3.70072","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149311984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}