{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T15:49:04Z","timestamp":1753890544509,"version":"3.41.2"},"reference-count":38,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T00:00:00Z","timestamp":1750291200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/2.zoppoz.workers.dev:443\/https\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Sci."],"abstract":"<jats:p>Mesenchymal stem cell therapy shows promising results for difficult-to-treat diseases, but manufacturing requires robust quality control through cell confluence monitoring. While deep learning can automate confluence estimation, research on cost-effective dataset curation and the role of foundation models in this task is limited. We investigate effective strategies for AI-based confluence estimation by studying active learning, goal-dependent labeling, and foundation models that require no training or labeling effort (zero-shot). Here, we show that zero-shot inference with the Segment Anything Model (SAM) achieves excellent confluence estimation without any task-specific training, outperforming even fine-tuned and specialized models. Moreover, our findings demonstrate that active learning does not significantly enhance training and performance compared to the random selection of training samples in homogeneous cell datasets. We demonstrate that streamlined labeling approaches tailored to specific goals yield results comparable to those of exhaustive, time-consuming annotation methods. Our results challenge common assumptions about dataset curation and model training: neither active learning nor extensive fine-tuning provided significant benefits in our real-world scenarios. Instead, we found that leveraging SAM's zero-shot capabilities and goal-dependent labeling offers the most cost-effective approach for AI-based confluence monitoring. Our study provides practical guidelines for implementing automated cell quality control in MSC manufacturing, demonstrating that extensive dataset curation may be unnecessary when foundation models can effectively handle the task right out of the box.<\/jats:p>","DOI":"10.3389\/fcomp.2025.1562358","type":"journal-article","created":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T05:30:58Z","timestamp":1750311058000},"update-policy":"https:\/\/2.zoppoz.workers.dev:443\/https\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Leveraging foundation models and goal-dependent annotations for automated cell confluence assessment"],"prefix":"10.3389","volume":"7","author":[{"given":"Maximilian Josef","family":"Joas","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Daniel","family":"Freund","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Robert","family":"Haase","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Erhard","family":"Rahm","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jan","family":"Ewald","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,6,19]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1101\/2023.08.21.554208","author":"Archit","year":"2023","journal-title":"Segment Anything for"},{"key":"B2","doi-asserted-by":"publisher","first-page":"777728","DOI":"10.3389\/fcomp.2022.777728","article-title":"LABKIT: Labeling and segmentation toolkit for big image data","volume":"4","author":"Arzt","year":"2022","journal-title":"Front. Comp. Sci"},{"key":"B3","doi-asserted-by":"publisher","first-page":"6204","DOI":"10.21037\/qims-24-801","article-title":"Instance segmentation of cells and nuclei from multi-organ cross-protocol microscopic images","volume":"14","author":"Baral","year":"2024","journal-title":"Quant. Imaging Med. Surg"},{"key":"B4","doi-asserted-by":"publisher","first-page":"679","DOI":"10.1109\/TPAMI.1986.4767851","article-title":"A computational approach to edge detection","volume":"8","author":"Canny","year":"1986","journal-title":"IEEE Trans. Pattern Analys. Mach. Intellig"},{"key":"B5","doi-asserted-by":"publisher","first-page":"6","DOI":"10.1091\/mbc.E22-08-0364","article-title":"Evaluation of cell segmentation methods without reference segmentations","volume":"34","author":"Chen","year":"2023","journal-title":"Mol. Biol. Cell"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1038","DOI":"10.1038\/s41592-021-01249-6","article-title":"LIVECell\u2014a large-scale dataset for label-free live cell segmentation","volume":"18","author":"Edlund","year":"2021","journal-title":"Nat. Methods"},{"key":"B7","doi-asserted-by":"publisher","first-page":"918565","DOI":"10.3389\/fimmu.2022.918565","article-title":"Optimization of mesenchymal stromal cell (MSC) manufacturing processes for a better therapeutic outcome","volume":"13","author":"Fern\u00e1ndez-Santos","year":"2022","journal-title":"Front. Immunol"},{"key":"B8","doi-asserted-by":"publisher","first-page":"824","DOI":"10.1016\/j.stem.2018.05.004","article-title":"Mesenchymal stromal cells: Clinical challenges and therapeutic opportunities","volume":"22","author":"Galipeau","year":"2018","journal-title":"Cell Stem Cell"},{"key":"B9","doi-asserted-by":"crossref","first-page":"467","DOI":"10.1007\/978-3-031-73116-7_27","article-title":"\u201cTRACKASTRA: transformer-based cell tracking for live-cell microscopy,\u201d","volume-title":"Computer Vision-ECCV 2024","author":"Gallusser","year":"2025"},{"key":"B10","doi-asserted-by":"publisher","first-page":"bbae282","DOI":"10.1093\/bib\/bbae282","article-title":"CryoSegNet: accurate cryo-EM protein particle picking by integrating the foundational AI image segmentation model and attention-gated U-Net","volume":"25","author":"Gyawali","year":"2024","journal-title":"Brief. Bioinform"},{"key":"B11","doi-asserted-by":"publisher","first-page":"225","DOI":"10.1016\/j.aiopen.2021.08.002","article-title":"Pre-trained models: Past, present and future","volume":"2","author":"Han","year":"2021","journal-title":"AI Open"},{"key":"B12","doi-asserted-by":"publisher","first-page":"220","DOI":"10.1016\/j.joen.2023.11.002","article-title":"Uncertainty-based active learning by bayesian U-Net for multi-label cone-beam CT segmentation","volume":"50","author":"Huang","year":"2024","journal-title":"J. Endod"},{"article-title":"\u201cGlobal burden of disease study 2019 (GBD 2019) disease and injury burden 1990-2019,\u201d","volume-title":"Technical report, Institute for Health Metrics and Evaluation (IHME)","year":"2020","key":"B13"},{"key":"B14","doi-asserted-by":"crossref","first-page":"7100","DOI":"10.1109\/ICASSP48485.2024.10447623","article-title":"\u201cThe power of few: Accelerating and enhancing data reweighting with coreset selection,\u201d","volume-title":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","author":"Jafari","year":"2024"},{"key":"B15","doi-asserted-by":"publisher","first-page":"2099","DOI":"10.1007\/s10278-024-01037-6","article-title":"Active learning in brain tumor segmentation with uncertainty sampling and annotation redundancy restriction","volume":"37","author":"Kim","year":"2024","journal-title":"J. Imag. Inform. Med"},{"key":"B16","doi-asserted-by":"publisher","first-page":"300","DOI":"10.3892\/br.2017.845","article-title":"Cell culture density affects the stemness gene expression of adipose tissue-derived mesenchymal stem cells","volume":"6","author":"Kim","year":"2017","journal-title":"Biomed. Reports"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41598-019-48525-2","article-title":"Burst statistics in an early biofilm quorum sensing model: the role of spatial colony-growth heterogeneity","volume":"9","author":"Kindler","year":"2019","journal-title":"Sci. Rep"},{"key":"B18","doi-asserted-by":"crossref","first-page":"3992","DOI":"10.1109\/ICCV51070.2023.00371","article-title":"\u201cSegment Anything,\u201d","volume-title":"2023 IEEE\/CVF International Conference on Computer Vision (ICCV)","author":"Kirillov","year":"2023"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1346","DOI":"10.1038\/s41551-022-00914-1","article-title":"Self-supervised learning in medicine and healthcare","volume":"6","author":"Krishnan","year":"2022","journal-title":"Nat. Biomed. Eng"},{"key":"B20","doi-asserted-by":"publisher","first-page":"102862","DOI":"10.1016\/j.media.2023.102862","article-title":"HAL-IA: A hybrid active learning framework using interactive annotation for medical image segmentation","volume":"88","author":"Li","year":"2023","journal-title":"Med. Image Anal"},{"key":"B21","doi-asserted-by":"crossref","first-page":"740","DOI":"10.1007\/978-3-319-10602-1_48","article-title":"\u201cMicrosoft COCO: Common objects in context,\u201d","volume-title":"Computer Vision-ECCV 2014","author":"Lin","year":"2014"},{"key":"B22","doi-asserted-by":"publisher","first-page":"947","DOI":"10.1109\/TMI.2022.3221666","article-title":"Which pixel to annotate: A label-efficient nuclei segmentation framework","volume":"42","author":"Lou","year":"2023","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B23","doi-asserted-by":"publisher","first-page":"50","DOI":"10.1214\/aoms\/1177730491","article-title":"On a test of whether one of two random variables is stochastically larger than the other","volume":"18","author":"Mann","year":"1947","journal-title":"Ann. Math. Statist"},{"volume-title":"Human-in-the-Loop Machine Learning: Active Learning and Annotation for Human-centered AI","year":"2021","author":"Monarch","key":"B24"},{"key":"B25","doi-asserted-by":"publisher","first-page":"2534","DOI":"10.1109\/TMI.2020.3048055","article-title":"Diminishing uncertainty within the training pool: active learning for medical image segmentation","volume":"40","author":"Nath","year":"2021","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B26","doi-asserted-by":"crossref","first-page":"2426","DOI":"10.1109\/ICIP.2018.8451852","article-title":"\u201cCell segmentation via region-based ellipse fitting,\u201d","volume-title":"2018 25th IEEE International Conference on Image Processing (ICIP)","author":"Panagiotakis","year":"2018"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2408.00714","article-title":"SAM 2: Segment anything in images and videos","author":"Ravi","year":"2024","journal-title":"arXiv"},{"key":"B28","first-page":"234","volume-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation","author":"Ronneberger","year":"2015"},{"key":"B29","doi-asserted-by":"publisher","first-page":"5283","DOI":"10.1007\/s10462-021-10021-3","article-title":"A review and experimental analysis of active learning over crowdsourced data","volume":"54","author":"Sayin","year":"2021","journal-title":"Artif. Intellig. Rev"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1523\/ENEURO.0056-18.2018","article-title":"Automatic cell segmentation by adaptive thresholding (ACSAT) for large-scale calcium imaging datasets","author":"Shen","year":"2018","journal-title":"eNeuro"},{"key":"B31","doi-asserted-by":"publisher","first-page":"1905","DOI":"10.3390\/cells12141905","article-title":"Management of rheumatoid arthritis: Possibilities and challenges of mesenchymal stromal\/stem cell-based therapies","volume":"12","author":"Shimizu","year":"2023","journal-title":"Cells"},{"key":"B32","doi-asserted-by":"publisher","first-page":"122647","DOI":"10.1016\/j.lfs.2024.122647","article-title":"Therapeutic applications of mesenchymal\/medicinal stem\/signaling cells preconditioned with external factors: are there more efficient approaches to utilize their regenerative potential?","volume":"346","author":"Strecanska","year":"2024","journal-title":"Life Sci"},{"key":"B33","doi-asserted-by":"publisher","first-page":"100","DOI":"10.1038\/s41592-020-01018-x","article-title":"Cellpose: a generalist algorithm for cellular segmentation","volume":"18","author":"Stringer","year":"2020","journal-title":"Nat. Methods"},{"key":"B34","doi-asserted-by":"publisher","first-page":"2276","DOI":"10.1038\/s41467-021-22518-0","article-title":"Democratising deep learning for microscopy with ZeroCostDL4Mic","volume":"12","author":"von Chamier","year":"2021","journal-title":"Nat. Commun"},{"key":"B35","unstructured":"Wu\n              Y.\n            \n            \n              Kirillov\n              A.\n            \n            \n              Massa\n              F.\n            \n            \n              Lo\n              W.-Y.\n            \n            \n              Girshick\n              R.\n            \n          \n          Detectron2\n          \n          2019"},{"key":"B36","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13321-023-00753-5","article-title":"Evaluating uncertainty-based active learning for accelerating the generalization of molecular property prediction","volume":"15","author":"Yin","year":"2023","journal-title":"J. Cheminform"},{"key":"B37","doi-asserted-by":"publisher","first-page":"2736","DOI":"10.1109\/TMI.2020.3046292","article-title":"Weakly supervised cell segmentation by point annotation","volume":"40","author":"Zhao","year":"2021","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B38","first-page":"1137","article-title":"\u201cActive learning with sampling by uncertainty and density for word sense disambiguation and text classification,\u201d","volume-title":"Proceedings of the 22nd International Conference on Computational Linguistics","author":"Zhu","year":"2008"}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/2.zoppoz.workers.dev:443\/https\/www.frontiersin.org\/articles\/10.3389\/fcomp.2025.1562358\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T05:31:02Z","timestamp":1750311062000},"score":1,"resource":{"primary":{"URL":"https:\/\/2.zoppoz.workers.dev:443\/https\/www.frontiersin.org\/articles\/10.3389\/fcomp.2025.1562358\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,19]]},"references-count":38,"alternative-id":["10.3389\/fcomp.2025.1562358"],"URL":"https:\/\/2.zoppoz.workers.dev:443\/https\/doi.org\/10.3389\/fcomp.2025.1562358","relation":{},"ISSN":["2624-9898"],"issn-type":[{"type":"electronic","value":"2624-9898"}],"subject":[],"published":{"date-parts":[[2025,6,19]]},"article-number":"1562358"}}