{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T06:02:12Z","timestamp":1789970532167,"version":"4.0.1"},"reference-count":61,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T00:00:00Z","timestamp":1789948800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Plant Sci."],"abstract":"<jats:p>\n                    Accurate and objective plant phenotyping is crucial for optimizing agricultural practices, understanding plant development, and enabling rapid responses to environmental changes. Traditional methods, often relying on visual observation, can be subjective, time-consuming, and may overlook subtle but important differences. This study demonstrates the power of combining digital imaging with deep learning to classify plant material with high accuracy, even when visual differences are minimal. We focused on differentiating between stage D (light green) and stage E (dark green) leaves of cacao (\n                    <jats:italic>Theobroma cacao L.<\/jats:italic>\n                    ), which are visually very similar in size and overall structure. Using cleared and stained leaves of the SCA 6 genotype to highlight the venation network, we trained a Vision Transformer (ViT) model, a deep learning architecture, on image patches. At the patch level, the model achieved an overall accuracy of 97.03% on an independent test set, with a recall of 96.0% for stage D and 98.3% for stage E. At the whole-leaf level, majority voting correctly classified 14 of 15 independent test leaves (93.3%). Attention maps indicated that image regions containing the midrib and primary lateral veins contributed strongly to classification. These attention maps identify discriminative image regions, but they do not by themselves determine the biological mechanism underlying the signal. The major-vein signal may reflect developmental differences in vein-associated structure, stage-associated differences in Safranin O uptake or optical density, tissue thickness, or a combination of these factors. Because vascular anatomy, lignification, hydraulic conductance, phloem loading, and source\u2013sink status were not directly measured, these mechanisms are treated as hypotheses requiring future anatomical, histochemical, and physiological validation. Thus, this study provides a proof-of-concept for interpretable image-based classification of stage D and stage E leaves within greenhouse-grown SCA 6 cacao. Extension to other cacao genotypes, field-grown plants, independent seasons, staining batches, stress detection, species identification, genotype discrimination, or precision-agriculture deployment will require external validation.\n                  <\/jats:p>","DOI":"10.3389\/fpls.2026.1885906","type":"journal-article","created":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T05:36:01Z","timestamp":1789968961000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Explainable growth stage classification of cacao (Theobroma cacao L.) leaves and key feature visualization using vision transformer and transfer learning"],"prefix":"10.3389","volume":"17","author":[{"given":"Ezekiel","family":"Ahn","sequence":"first","affiliation":[{"name":"Sustainable Perennial Crops Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eun-Sung","family":"Park","sequence":"additional","affiliation":[{"name":"Department of Smart Agricultural Systems, Chungnam National University","place":["Daejeon, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Moon S.","family":"Kim","sequence":"additional","affiliation":[{"name":"Environmental Microbial and Food Safety Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hangi","family":"Kim","sequence":"additional","affiliation":[{"name":"Department of Biosystems Machinery Engineering, Chungnam National University","place":["Daejeon, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lalit M.","family":"Kandpal","sequence":"additional","affiliation":[{"name":"Environmental Microbial and Food Safety Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sunchung","family":"Park","sequence":"additional","affiliation":[{"name":"Sustainable Perennial Crops Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Seunghyun","family":"Lim","sequence":"additional","affiliation":[{"name":"Sustainable Perennial Crops Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lyndel W.","family":"Meinhardt","sequence":"additional","affiliation":[{"name":"Sustainable Perennial Crops Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Byoung-Kwan","family":"Cho","sequence":"additional","affiliation":[{"name":"Department of Smart Agricultural Systems, Chungnam National University","place":["Daejeon, Republic of Korea"]},{"name":"Department of Biosystems Machinery Engineering, Chungnam National University","place":["Daejeon, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Insuck","family":"Baek","sequence":"additional","affiliation":[{"name":"Environmental Microbial and Food Safety Laboratory, Agricultural Research Service, United States Department of Agriculture","place":["Beltsville, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,21]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1007\/s10343-022-00796-y","article-title":"Improving deep learning-based plant disease classification with attention mechanism","volume":"75","author":"Alirezazadeh","year":"2023","journal-title":"Gesunde Pflanzen"},{"key":"B2","doi-asserted-by":"publisher","first-page":"425","DOI":"10.1590\/s1677-04202007000400011","article-title":"Ecophysiology of the cacao tree","volume":"19","author":"Almeida","year":"2007","journal-title":"Braz. J. Plant Physiol."},{"key":"B3","doi-asserted-by":"publisher","first-page":"256","DOI":"10.3390\/sym11020256","article-title":"Identification and classification of maize drought stress using deep convolutional neural network","volume":"11","author":"An","year":"2019","journal-title":"Symmetry"},{"key":"B4","doi-asserted-by":"publisher","first-page":"108650","DOI":"10.1016\/j.measurement.2020.108650","article-title":"A deep learning approach to measure stress level in plants due to nitrogen deficiency","volume":"173","author":"Azimi","year":"2021","journal-title":"Measurement"},{"key":"B5","doi-asserted-by":"publisher","first-page":"554","DOI":"10.1038\/s42003-025-08019-6","article-title":"Spatial patterning of chloroplasts and stomata in developing cacao leaves","volume":"8","author":"Baek","year":"2025","journal-title":"Commun. Biol."},{"key":"B6","doi-asserted-by":"publisher","first-page":"1247","DOI":"10.1016\/j.plantsci.2005.01.002","article-title":"Gene expression in leaves of Theobroma cacao in response to mechanical wounding, ethylene, and\/or methyl jasmonate","volume":"168","author":"Bailey","year":"2005","journal-title":"Plant Sci."},{"key":"B7","doi-asserted-by":"publisher","first-page":"1315","DOI":"10.1111\/j.1469-8137.1973.tb02109.x","article-title":"Biochemical and physiological aspects of leaf development in cocoa (Theobroma cacao) I. Development of chlorophyll and photosynthetic activity","volume":"72","author":"Baker","year":"1973","journal-title":"New Phytol."},{"key":"B8","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-030-23265-8_12","article-title":"Cacao (Theobroma cacao L.) breeding","volume-title":"Advances in Plant Breeding Strategies: Industrial and Food Crops: Volume 6","author":"Bekele","year":"2019"},{"key":"B9","doi-asserted-by":"publisher","first-page":"635","DOI":"10.1016\/s0168-9452(03)00018-9","article-title":"Photoinhibition of photosynthesis in mature and young leaves of grapevine (Vitis vinifera L.)","volume":"164","author":"Bertamini","year":"2003","journal-title":"Plant Sci."},{"key":"B10","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1146\/annurev-genet-102209-163508","article-title":"The genetics of lignin biosynthesis: Connecting genotype to phenotype","volume":"44","author":"Bonawitz","year":"2010","journal-title":"Annu. Rev. Genet."},{"key":"B11","doi-asserted-by":"publisher","first-page":"1890","DOI":"10.1104\/pp.107.101352","article-title":"Leaf maximum photosynthetic rate and venation are linked by hydraulics","volume":"144","author":"Brodribb","year":"2007","journal-title":"Plant Physiol."},{"key":"B12","doi-asserted-by":"publisher","first-page":"105393","DOI":"10.1016\/j.compag.2020.105393","article-title":"Using deep transfer learning for image-based plant disease identification","volume":"173","author":"Chen","year":"2020","journal-title":"Comput. Electron. Agric."},{"key":"B13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.plaphy.2013.05.009","article-title":"Plant phenolics: Recent advances on their biosynthesis, genetics, and ecophysiology","volume":"72","author":"Cheynier","year":"2013","journal-title":"Plant Physiol. Biochem."},{"key":"B14","volume-title":"Deep Learning for Precision Agriculture: Evaluating Cnns and Vision Transformers in Rice Disease Classification","author":"Chitta","year":"2024"},{"key":"B15","doi-asserted-by":"publisher","first-page":"2275","DOI":"10.1007\/s11277-024-11374-y","article-title":"Classification of different plant species using deep learning and machine learning algorithms","volume":"136","author":"Chouhan","year":"2024","journal-title":"Wireless Pers. Commun."},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.3389\/fgene.2022.822173","article-title":"Plant genotype to phenotype prediction using machine learning","volume":"13","author":"Danilevicz","year":"2022","journal-title":"Front. Genet."},{"key":"B17","volume-title":"Livelihood Strategies and Risk Behavior of Cacao Producers in Ecuador: Effects of National Policies to Support Cacao Farmers and Specialty Cacao Landraces","author":"D\u00edaz-Montenegro","year":"2019"},{"key":"B18","doi-asserted-by":"publisher","first-page":"859","DOI":"10.1016\/S0031-9422(01)00049-8","article-title":"Lignification and lignin topochemistry\u2014an ultrastructural view","volume":"57","author":"Donaldson","year":"2001","journal-title":"Phytochemistry"},{"key":"B19","volume-title":"An Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale","author":"Dosovitskiy","year":"2020"},{"key":"B20","doi-asserted-by":"publisher","first-page":"517","DOI":"10.3832\/ifor3936-014","article-title":"Estimation of canopy attributes of wild cacao trees using digital cover photography and machine learning algorithms","volume":"14","author":"Duarte-Carvajalino","year":"2021","journal-title":"iForest  Biogeosc. For."},{"key":"B21","doi-asserted-by":"crossref","DOI":"10.1109\/ICEENG58856.2024.10566412","volume-title":"Deep Learning Based Grapevine Leaf Classification Using Augmented Images and Multi-Classifier Fusion for Improved Accuracy and Precision","author":"Elkassar","year":"2024"},{"key":"B22","doi-asserted-by":"publisher","first-page":"1540","DOI":"10.30574\/ijsra.2024.11.1.0186","article-title":"A review of machine learning applications in cacao post-harvest management","volume":"11","author":"Ferran","year":"2024","journal-title":"Int. J. Sci. Res. Arch."},{"key":"B23","doi-asserted-by":"publisher","first-page":"3054","DOI":"10.3390\/s19143054","article-title":"Spatial variability of aroma profiles of cocoa trees obtained through computer vision and machine learning modelling: A cover photography and high spatial remote sensing application","volume":"19","author":"Fuentes","year":"2019","journal-title":"Sensors"},{"key":"B24","doi-asserted-by":"crossref","DOI":"10.1109\/STSIVA.2019.8730258","volume-title":"Predictive Model for Cocoa Yield in Santander Using Supervised Machine Learning","author":"Gamboa","year":"2019"},{"key":"B25","doi-asserted-by":"publisher","first-page":"29","DOI":"10.3390\/agriengineering2030029","article-title":"Deep learning application in plant stress imaging: A review","volume":"2","author":"Gao","year":"2020","journal-title":"AgriEngineering"},{"key":"B26","doi-asserted-by":"crossref","DOI":"10.1002\/9781118350089","volume-title":"Color in Computer Vision: Fundamentals and Applications","author":"Gevers","year":"2012"},{"key":"B27","volume-title":"Applying a One Health Approach to Study the Livelihoods of Cocoa Farming Communities in Bougainville","author":"Hall","year":"2022"},{"key":"B28","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1109\/tpami.2022.3152247","article-title":"A survey on vision transformer","volume":"45","author":"Han","year":"2022","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"B29","doi-asserted-by":"publisher","first-page":"4015","DOI":"10.3390\/plants12234015","article-title":"Plant physiological analysis to overcome limitations to plant phenotyping","volume":"12","author":"Haworth","year":"2023","journal-title":"Plants"},{"key":"B30","volume-title":"Computer-Vision Benchmark Segment-Anything Model (Sam) in Medical Images: Accuracy in 12 Datasets","author":"He","year":"2023"},{"key":"B31","doi-asserted-by":"publisher","first-page":"311","DOI":"10.15446\/acag.v67n2.63046","article-title":"Quality of cocoa (Theobroma cacao L.) DNA from foliar tissue at different stages of development","volume":"67","author":"Henao Ram\u00edrez","year":"2018","journal-title":"Acta Agron\u00f3mica"},{"key":"B32","doi-asserted-by":"publisher","first-page":"259","DOI":"10.1016\/j.neucom.2022.06.083","article-title":"Recent advances on image edge detection: A comprehensive review","volume":"503","author":"Jing","year":"2022","journal-title":"Neurocomputing"},{"key":"B33","volume-title":"Plant Microtechnique","author":"Johansen","year":"1940"},{"key":"B34","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV51070.2023.00371","volume-title":"Segment Anything","author":"Kirillov","year":"2023"},{"key":"B35","doi-asserted-by":"publisher","first-page":"108885","DOI":"10.1016\/j.jneumeth.2020.108885","article-title":"Data augmentation for deep-learning-based electroencephalography","volume":"346","author":"Lashgari","year":"2020","journal-title":"J. Neurosci. Methods"},{"key":"B36","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.patcog.2017.05.015","article-title":"How deep learning extracts and learns leaf features for plant classification","volume":"71","author":"Lee","year":"2017","journal-title":"Pattern Recognit."},{"key":"B37","doi-asserted-by":"publisher","first-page":"e42407","DOI":"10.1016\/j.heliyon.2025.e42407","article-title":"Cacao floral traits are shaped by the interaction of flower position with genotype","volume":"11","author":"Lim","year":"2025","journal-title":"Heliyon"},{"key":"B38","doi-asserted-by":"publisher","first-page":"1077","DOI":"10.1007\/s10457-022-00765-4","article-title":"Contribution of cacao agroforestry versus mono-cropping systems for enhanced sustainability. A review with a focus on yield","volume":"96","author":"Mattalia","year":"2022","journal-title":"Agrofor. Syst."},{"key":"B39","doi-asserted-by":"publisher","first-page":"379","DOI":"10.1021\/bk-2012-1095.ch018","article-title":"Expression of designed antimicrobial peptides in Theobroma cacao L. trees reduces leaf necrosis caused by Phytophthora spp","volume":"1095","author":"Mej\u00eda","year":"2012","journal-title":"Small Wonders: Peptides For. Dis. Control"},{"key":"B40","doi-asserted-by":"publisher","first-page":"104859","DOI":"10.1016\/j.compag.2019.104859","article-title":"Crop yield prediction with deep convolutional neural networks","volume":"163","author":"Nevavuori","year":"2019","journal-title":"Comput. Electron. Agric."},{"key":"B41","doi-asserted-by":"publisher","first-page":"6949","DOI":"10.3390\/s23156949","article-title":"Visual intelligence in precision agriculture: Exploring plant disease detection via efficient vision transformers","volume":"23","author":"Parez","year":"2023","journal-title":"Sensors"},{"key":"B42","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1111\/j.1461-0248.2011.01712.x","article-title":"Scaling and structure of dicotyledonous leaf venation networks","volume":"15","author":"Price","year":"2012","journal-title":"Ecol. Lett."},{"key":"B43","volume-title":"Imagenet-21k Pretraining for the Masses","author":"Ridnik","year":"2021"},{"key":"B44","doi-asserted-by":"publisher","first-page":"553","DOI":"10.1006\/anbo.2001.1391","article-title":"Evolution and function of leaf venation architecture: A review","volume":"87","author":"Roth-Nebelsick","year":"2001","journal-title":"Ann. Bot."},{"key":"B45","doi-asserted-by":"publisher","first-page":"983","DOI":"10.1111\/nph.12253","article-title":"Leaf venation: Structure, function, development, evolution, ecology and applications in the past, present and future","volume":"198","author":"Sack","year":"2013","journal-title":"New Phytol."},{"key":"B46","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13007-020-00647-9","article-title":"Deep learning-based detection of seedling development","volume":"16","author":"Samiei","year":"2020","journal-title":"Plant Methods"},{"key":"B47","doi-asserted-by":"publisher","first-page":"197","DOI":"10.1016\/j.media.2019.01.012","article-title":"Attention gated networks: Learning to leverage salient regions in medical images","volume":"53","author":"Schlemper","year":"2019","journal-title":"Med. Image Anal."},{"key":"B48","doi-asserted-by":"publisher","first-page":"3796","DOI":"10.3390\/s19173796","article-title":"Precision agriculture techniques and practices: From considerations to applications","volume":"19","author":"Shafi","year":"2019","journal-title":"Sensors"},{"key":"B49","doi-asserted-by":"publisher","first-page":"585","DOI":"10.1080\/0952813x.2022.2096698","article-title":"Cognitive framework and learning paradigms of plant leaf classification using artificial neural network and support vector machine","volume":"36","author":"Sharma","year":"2024","journal-title":"J. Exp. Theor. Artif. Intell."},{"key":"B50","doi-asserted-by":"publisher","first-page":"286","DOI":"10.37391\/ijeer.120139","article-title":"Nutrient deficiency of paddy leaf classification using hybrid convolutional neural network","volume":"12","author":"Sherline Jesie","year":"2024","journal-title":"Int. J. Electrical Electron. Res."},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.1111\/jfpe.13175","article-title":"Sensing fermentation degree of cocoa (Theobroma cacao L.) beans by machine learning classification models based electronic nose system","volume":"42","author":"Tan","year":"2019","journal-title":"J. Food Process Eng."},{"key":"B52","doi-asserted-by":"publisher","first-page":"1580","DOI":"10.3390\/s18051580","article-title":"Weed growth stage estimator using deep convolutional neural networks","volume":"18","author":"Teimouri","year":"2018","journal-title":"Sensors"},{"key":"B53","doi-asserted-by":"publisher","first-page":"9237136","DOI":"10.1155\/2019\/9237136","article-title":"How convolutional neural networks diagnose plant disease","volume":"2019","author":"Toda","year":"2019","journal-title":"Plant Phenomics"},{"key":"B54","doi-asserted-by":"publisher","first-page":"225","DOI":"10.46604\/ijeti.2022.8865","article-title":"Deep learning for image-based plant growth monitoring: a review","volume":"12","author":"Tong","year":"2022","journal-title":"Int. J. Eng. Technol. Innovation"},{"key":"B55","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1146\/annurev.pp.40.060189.001003","article-title":"The sink-source transition in leaves","volume":"40","author":"Turgeon","year":"1989","journal-title":"Annual Review of Plant Physiology and Plant Molecular Biology"},{"key":"B56","doi-asserted-by":"publisher","first-page":"1400038","DOI":"10.3732\/apps.1400038","article-title":"Preparation of samples for leaf architecture studies, a method for mounting cleared leaves","volume":"2","author":"Vasco","year":"2014","journal-title":"Appl. Plant Sci."},{"key":"B57","first-page":"96","article-title":"On the usefulness of attention for object recognition","volume":"3368","author":"Walther","year":"2005","journal-title":"Lecture Notes in Computer Science"},{"key":"B58","doi-asserted-by":"publisher","first-page":"114770","DOI":"10.1016\/j.eswa.2021.114770","article-title":"Rice diseases detection and classification using attention based neural network and bayesian optimization","volume":"178","author":"Wang","year":"2021","journal-title":"Expert Syst. Appl."},{"key":"B59","doi-asserted-by":"publisher","first-page":"224","DOI":"10.1016\/j.cviu.2010.10.002","article-title":"A survey of vision-based methods for action representation, segmentation and recognition","volume":"115","author":"Weinland","year":"2011","journal-title":"Comput. Vision Image Understanding"},{"key":"B60","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.njas.2015.09.001","article-title":"Cocoa production in West Africa, a review and analysis of recent developments","volume":"74","author":"Wessel","year":"2015","journal-title":"NJAS Wageningen J. Life Sci."},{"key":"B61","doi-asserted-by":"publisher","first-page":"4","DOI":"10.1111\/pbi.12848","article-title":"Cacao biotechnology: Current status and future prospects","volume":"16","author":"Wickramasuriya","year":"2018","journal-title":"Plant Biotechnol. J."}],"container-title":["Frontiers in Plant Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpls.2026.1885906\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T05:36:05Z","timestamp":1789968965000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpls.2026.1885906\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,21]]},"references-count":61,"alternative-id":["10.3389\/fpls.2026.1885906"],"URL":"https:\/\/doi.org\/10.3389\/fpls.2026.1885906","relation":{},"ISSN":["1664-462X"],"issn-type":[{"value":"1664-462X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,21]]},"article-number":"1885906"}}