{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:51:58Z","timestamp":1791525118852,"version":"4.3.4"},"reference-count":37,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T00:00:00Z","timestamp":1791504000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100003627","name":"Rural Development Administration","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100003627","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Plant Sci."],"abstract":"\n Introduction<\/jats:title>\n Morphological characterization of agricultural genetic resources relies largely on visual assessment by experts, and applying the same criteria repeatedly to large collections is labor-intensive. Image-based deep learning can support this process, but systematic evidence on training strategies suited to genebank trait data remains limited.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n This study targeted six morphological traits of lettuce genetic resources characterized according to the test guidelines of the International Union for the Protection of New Varieties of Plants (UPOV). Using genebank images accumulated from 2012 to 2025, six convolutional neural networks were trained with three training strategies under identical conditions, and their performance was compared.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n Fine-tuning achieved the highest mean macro F1-score in 35 of the 36 backbone\u2013trait combinations and significantly outperformed training from scratch and feature extraction, whereas no single backbone was superior across all traits. Misclassifications were concentrated in morphologically adjacent classes rather than randomly distributed, and for the ordinal traits, errors spanning two or more class levels were rare.<\/jats:p>\n <\/jats:sec>\n \n Discussion<\/jats:title>\n These patterns suggest that the remaining classification difficulty was associated in part with the morphological continuity between neighboring classes. The attainable performance also differed among traits, distinguishing traits for which the training strategy alone achieved relatively high performance from those for which it did not. These results provide empirical evidence for selecting training strategies and for determining the role of model-assisted characterization on a trait-by-trait basis in genebanks.<\/jats:p>\n <\/jats:sec>","DOI":"10.3389\/fpls.2026.1970840","type":"journal-article","created":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:33:19Z","timestamp":1791523999000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Fine-tuning improves deep learning-based digital phenotyping of lettuce genetic resources"],"prefix":"10.3389","volume":"17","author":[{"given":"Eunju","family":"Lee","sequence":"first","affiliation":[{"name":"Department of Agricultural Engineering, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuseok","family":"Jeong","sequence":"additional","affiliation":[{"name":"Department of Agricultural Engineering, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hyo Jun","family":"Bae","sequence":"additional","affiliation":[{"name":"Department of Agricultural Engineering, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yoon-Jung","family":"Lee","sequence":"additional","affiliation":[{"name":"National Agrobiodiversity Center, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Seong-Hoon","family":"Kim","sequence":"additional","affiliation":[{"name":"National Agrobiodiversity Center, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Young-Wang","family":"Na","sequence":"additional","affiliation":[{"name":"National Agrobiodiversity Center, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hwan-suk","family":"Cheong","sequence":"additional","affiliation":[{"name":"Agricultural AI and Data Division, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Inchan","family":"Choi","sequence":"additional","affiliation":[{"name":"Department of Agricultural Engineering, National Institute of Agricultural Sciences, Rural Development Administration","place":["Jeonju, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,10,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.3389\/fpls.2025.1679498","article-title":"Genetic variation in storage protein and starch biosynthesis genes reveals key associations with seed composition in pea (pisum sativum)","volume":"16","author":"Alemu","year":"2025","journal-title":"Front. Plant Sci."},{"key":"B2","doi-asserted-by":"publisher","first-page":"104874","DOI":"10.1016\/j.compag.2019.104874","article-title":"Identification of haploid and diploid maize seeds using convolutional neural networks and a transfer learning approach","volume":"163","author":"Altunta\u015f","year":"2019","journal-title":"Comput. Electron. Agric."},{"key":"B3","doi-asserted-by":"publisher","first-page":"961","DOI":"10.1007\/s10460-023-10417-7","article-title":"Genebanking plant genetic resources in the postgenomic era","volume":"40","author":"Aubry","year":"2023","journal-title":"Agric. Hum. Values"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1111\/ppl.14482","article-title":"Genetic mosaic of the mediterranean fig: Comprehensive genomic insights from a gene bank collection","volume":"176","author":"Bazakos","year":"2024","journal-title":"Physiol. Plant"},{"key":"B5","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1016\/j.neunet.2018.07.011","article-title":"A systematic study of the class imbalance problem in convolutional neural networks","volume":"106","author":"Buda","year":"2018","journal-title":"Neural Networks"},{"key":"B6","doi-asserted-by":"publisher","first-page":"213","DOI":"10.1037\/h0026256","article-title":"Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit","volume":"70","author":"Cohen","year":"1968","journal-title":"psychol. Bull."},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0302158","article-title":"Digital descriptors sharpen classical descriptors, for improving genebank accession management: A case study on arachis spp. And phaseolus spp","volume":"19","author":"Conejo-Rodr\u00edguez","year":"2024","journal-title":"PloS One"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.3390\/math13152539","article-title":"An analysis of layer-freezing strategies for enhanced transfer learning in YOLO architectures","volume":"13","author":"Dobrzycki","year":"2025","journal-title":"Mathematics"},{"key":"B9","doi-asserted-by":"publisher","first-page":"23","DOI":"10.1002\/ppp3.10570","article-title":"Realizing the potential of plant genetic resources: The use of phenomics for genebanks","volume":"7","author":"Ghamkhar","year":"2025","journal-title":"Plants People Planet"},{"key":"B10","doi-asserted-by":"publisher","first-page":"228","DOI":"10.1016\/j.neucom.2017.01.018","article-title":"Plant identification using deep neural networks via optimization of transfer learning parameters","volume":"235","author":"Ghazi","year":"2017","journal-title":"Neurocomputing"},{"key":"B11","doi-asserted-by":"publisher","first-page":"1407","DOI":"10.1111\/nph.14993","article-title":"Plant genetic resources for food and agriculture: Opportunities and challenges emerging from the science and information technology revolution","volume":"217","author":"Halewood","year":"2018","journal-title":"New Phytol."},{"key":"B12","first-page":"770","article-title":"Deep residual learning for image recognition","author":"He","year":"2016"},{"key":"B13","doi-asserted-by":"publisher","first-page":"97","DOI":"10.1007\/s10462-024-11081-x","article-title":"Transfer learning in agriculture: A review","volume":"58","author":"Hossen","year":"2025","journal-title":"Artif. Intell. Rev."},{"key":"B14","first-page":"4700","article-title":"Densely connected convolutional networks","author":"Huang","year":"2017"},{"key":"B15","doi-asserted-by":"publisher","first-page":"88","DOI":"10.1002\/csc2.21135","article-title":"Characterization of agronomic traits and genomic diversity in a newly assembled radish core collection","volume":"64","author":"Huh","year":"2024","journal-title":"Crop Sci."},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.7717\/peerj.15043","article-title":"Evaluation of a dill (anethum graveolens l.) gene bank germplasm collection using multivariate analysis of morphological traits, molecular genotyping and chemical composition to identify novel genotypes for plant breeding","volume":"11","author":"Kadoglidou","year":"2023","journal-title":"PeerJ"},{"key":"B17","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1016\/j.compag.2019.01.041","article-title":"Analysis of transfer learning for deep neural network based plant classification models","volume":"158","author":"Kaya","year":"2019","journal-title":"Comput. Electron. Agric."},{"key":"B18","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1186\/s13007-021-00787-6","article-title":"DeepCob: Precise and high-throughput analysis of maize cob geometry using deep learning with an application in genebank phenomics","volume":"17","author":"Kienbaum","year":"2021","journal-title":"Plant Methods"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.3390\/plants12152825","article-title":"RDA-genebank and digital phenotyping for next-generation research on plant genetic resources","volume":"12","author":"Kim","year":"2023","journal-title":"Plants"},{"key":"B20","doi-asserted-by":"publisher","first-page":"566","DOI":"10.1038\/s41597-026-06908-x","article-title":"Wheat historical phenotypic data from european genebanks as an important resource for research and breeding","volume":"13","author":"Le Floch","year":"2026","journal-title":"Sci. Data"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1016\/j.compag.2020.105672","article-title":"A review of computer vision technologies for plant phenotyping","volume":"176","author":"Li","year":"2020","journal-title":"Comput. Electron. Agric."},{"key":"B22","first-page":"11976","article-title":"A convnet for the 2020s","author":"Liu","year":"2022"},{"key":"B23","doi-asserted-by":"publisher","first-page":"329","DOI":"10.1007\/s10681-012-0750-9","article-title":"Comparison of image analysis and direct measurement of UPOV taxonomic characteristics for variety discrimination as determined over five growing seasons, using industrial chicory as a model crop","volume":"189","author":"Lootens","year":"2013","journal-title":"Euphytica"},{"key":"B24","doi-asserted-by":"publisher","first-page":"3183","DOI":"10.1016\/j.csbj.2023.05.005","article-title":"Machine learning classification of plant genotypes grown under different light conditions through the integration of multi-scale time-series data","volume":"21","author":"Sakeef","year":"2023","journal-title":"Comput. Struct. Biotechnol. J."},{"key":"B25","first-page":"4510","article-title":"Mobilenetv2: Inverted residuals and linear bottlenecks","author":"Sandler","year":"2018"},{"key":"B26","first-page":"618","article-title":"Grad-cam: Visual explanations from deep networks via gradient-based localization","author":"Selvaraju","year":"2017"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1409.1556","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014","journal-title":"arXiv Preprint arXiv:1409.1556"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.3389\/fpls.2026.1790903","article-title":"Deep learning techniques for early detection and classification of leaf diseases in crops","volume":"17","author":"Srinivasu","year":"2026","journal-title":"Front. Plant Sci."},{"key":"B29","doi-asserted-by":"publisher","first-page":"66","DOI":"10.1186\/s13007-018-0333-4","article-title":"Deep phenotyping: Deep learning for temporal phenotype\/genotype classification","volume":"14","author":"Taghavi Namin","year":"2018","journal-title":"Plant Methods"},{"key":"B30","first-page":"6105","article-title":"Efficientnet: Rethinking model scaling for convolutional neural networks","author":"Tan","year":"2019"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1016\/j.compag.2025.110149","article-title":"LGENetB4CA: A novel deep learning approach for chili germplasm differentiation and leaf disease classification","volume":"233","author":"Van","year":"2025","journal-title":"Comput. Electron. Agric."},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.3390\/plants10112260","article-title":"Integrating genomic and phenomic approaches to support plant genetic resources conservation and use","volume":"10","author":"Volk","year":"2021","journal-title":"Plants"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.3389\/fpls.2025.1732979","article-title":"Artificial intelligence in plant science: From image-based phenotyping to yield and trait prediction","volume":"16","author":"Wang","year":"2026","journal-title":"Front. Plant Sci."},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.3390\/plants9081050","article-title":"Document or lose it\u2014on the importance of information management for genetic resources conservation in genebanks","volume":"9","author":"Weise","year":"2020","journal-title":"Plants"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1016\/j.compag.2021.106679","article-title":"Study of chrysanthemum image phenotype on-line classification based on transfer learning and bilinear convolutional neural network","volume":"194","author":"Yuan","year":"2022","journal-title":"Comput. Electron. Agric."},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1016\/j.compag.2022.106930","article-title":"A novel image detection method for internal cracks in corn seeds in an industrial inspection line","volume":"197","author":"Zhang","year":"2022","journal-title":"Comput. Electron. Agric."},{"key":"B37","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1109\/JPROC.2020.3004555","article-title":"A comprehensive survey on transfer learning","volume":"109","author":"Zhuang","year":"2020","journal-title":"Proc. IEEE"}],"container-title":["Frontiers in Plant Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpls.2026.1970840\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:33:20Z","timestamp":1791524000000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpls.2026.1970840\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,9]]},"references-count":37,"alternative-id":["10.3389\/fpls.2026.1970840"],"URL":"https:\/\/doi.org\/10.3389\/fpls.2026.1970840","relation":{},"ISSN":["1664-462X"],"issn-type":[{"value":"1664-462X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,9]]},"article-number":"1970840"}}