{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T06:15:49Z","timestamp":1783577749476,"version":"3.55.0"},"reference-count":47,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T00:00:00Z","timestamp":1783555200000},"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. Bioeng. Biotechnol."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Monitoring microcarrier colonization during mesenchymal stromal cell (MSC) expansion in bioreactors is essential for process control. Current manual quantification methods are time-consuming, subjective, and prone to inter-operator variability. This study presents a proof-of-concept evaluation of deep learning-based object detection for automated microcarrier colonization quantification in bioreactor cell cultures<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>Three state-of-the-art architectures\u2014YOLOv8, YOLOv11, and RT-DETR\u2014were compared within a unified experimental framework. A dataset of 699 fluorescence microscopy images containing 46,982 annotated microcarriers (classified as colonized or non-colonized) was used for training and evaluation. Model selection followed a three-stage Successive Halving Algorithm (SHA) strategy to efficiently identify the best-performing architecture across 12 configurations without requiring high-end computational infrastructure. COCO-pretrained transfer learning was applied across all models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>YOLOv8-l emerged as the top-performing model, achieving mAP50\u201395 of 0.855 on validation and 0.787 on test, with a colonization estimation error (Mean Absolute Error) of 11.75%. The dataset size proved sufficient for CNN-based architectures, while also revealing the higher data demands of transformer-based detectors.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These results demonstrate that automated deep learning pipelines can reliably quantify microcarrier colonization, offering a practical alternative to manual assessment in cell therapy manufacturing. The findings also suggest that newer YOLO models do not necessarily yield improvements in domain-specific biomedical tasks<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fbioe.2026.1849366","type":"journal-article","created":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T05:36:40Z","timestamp":1783575400000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Comparative evaluation of YOLOv8, YOLOv11, and RT-DETR for automated microcarrier colonization assessment in bioreactor cell cultures"],"prefix":"10.3389","volume":"14","author":[{"given":"Manuel T.","family":"Fern\u00e1ndez de Sevilla","sequence":"first","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ra\u00fal","family":"Valero","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Beatriz","family":"Men\u00e9ndez","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gonzalo","family":"Garc\u00eda Dom\u00ednguez","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cristina","family":"Blanco S\u00e1nchez","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Natalia","family":"Canales","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ram\u00f3n","family":"Menta Hern\u00e1ndez","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marta H. G.","family":"Costa","sequence":"additional","affiliation":[{"name":"iBET - Instituto de Biologia Experimental e Tecnol\u00f3gica","place":["Oeiras, Portugal"]},{"name":"Instituto de Tecnologia Qu\u00edmica e Biol\u00f3gica Ant\u00f3nio Xavier, Universidade Nova de Lisboa","place":["Oeiras, Portugal"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Margarida","family":"Serra","sequence":"additional","affiliation":[{"name":"iBET - Instituto de Biologia Experimental e Tecnol\u00f3gica","place":["Oeiras, Portugal"]},{"name":"Instituto de Tecnologia Qu\u00edmica e Biol\u00f3gica Ant\u00f3nio Xavier, Universidade Nova de Lisboa","place":["Oeiras, Portugal"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eleuterio","family":"Lombardo","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alvaro","family":"Avivar\u2010Valderas","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pablo","family":"Manche\u00f1o-Corvo","sequence":"additional","affiliation":[{"name":"Cell Therapies Science Department, Takeda Madrid Cell Therapy Technology Center","place":["Madrid, Spain"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"153","DOI":"10.4236\/jcc.2024.1210011","article-title":"Defect detection in manufacturing: an integrated deep learning approach","volume":"12","author":"Adeyemi","year":"2024","journal-title":"J. Comput. Commun."},{"key":"B2","doi-asserted-by":"publisher","first-page":"310","DOI":"10.22088\/IJMCM.BUMS.12.3.310","article-title":"Stem cell therapy, the market, the opportunities and the threat","volume":"12","author":"Bahari","year":"2023","journal-title":"Ijmcmed"},{"key":"B3","unstructured":"YOLOv4: optimal speed and accuracy of object detection\n          \n          \n            \n              Bochkovskiy\n              A.\n            \n            \n              Wang\n              C.-Y.\n            \n            \n              Liao\n              H.-Y. M.\n            \n          \n          \n          2020"},{"key":"B4","first-page":"30","article-title":"Cell therapy bioprocessing integrating process and product development for the next generation of biotherapeutics","volume":"9","author":"Brandenberger","year":"2011","journal-title":"Bioprocess Int."},{"key":"B5","doi-asserted-by":"publisher","first-page":"201","DOI":"10.1016\/j.jcyt.2023.11.008","article-title":"Assessment and comparison of viability assays for cellular products","volume":"26","author":"Cai","year":"2024","journal-title":"Cytotherapy"},{"key":"B6","doi-asserted-by":"publisher","first-page":"e0148469","DOI":"10.1371\/journal.pone.0148469","article-title":"High-throughput method for automated colony and cell counting by digital image analysis based on edge detection","volume":"11","author":"Choudhry","year":"2016","journal-title":"PLOS ONE"},{"key":"B7","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1186\/s12896-026-01101-9","article-title":"Scaling the manufacture of adipose tissue-derived mesenchymal stromal cells: integrated bioreactor workflows from inoculation to harvest without the need of seed train","volume":"26","author":"Costa","year":"2026","journal-title":"BMC Biotechnol."},{"key":"B8","unstructured":"RandAugment: practical automated data augmentation with a reduced search space\n          \n          \n            \n              Cubuk\n              E. D.\n            \n            \n              Zoph\n              B.\n            \n            \n              Shlens\n              J.\n            \n            \n              Le\n              Q. V.\n            \n          \n          \n          2019"},{"key":"B9","doi-asserted-by":"crossref","DOI":"10.1016\/j.msec.2019.109782","article-title":"Attachment and detachment strategies in microcarrier-based cell culture technology: a comprehensive review","volume-title":"Materials Science and Engineering C","author":"Derakhti","year":"2019"},{"key":"B10","first-page":"199","article-title":"Bioreactor expansion of human mesenchymal stem cells according to GMP requirements","volume-title":"Stem Cells and Good Manufacturing Practices: Methods, Protocols, and Regulations","author":"Elseberg","year":"2015"},{"key":"B11","unstructured":"Analytical procedures and methods validation for drugs and biologics guidance for industry\n          \n          \n          2015"},{"key":"B12","doi-asserted-by":"publisher","first-page":"105717","DOI":"10.1016\/j.ejps.2021.105717","article-title":"Applications of machine vision in pharmaceutical technology: a review","volume":"159","author":"Galata","year":"2021","journal-title":"Eur. J. Pharm. Sci."},{"key":"B13","unstructured":"YOLOX: exceeding YOLO series in 2021 V100 batch 1 latency (ms) YOLOX-L YOLOv5-L YOLOX-DarkNet53 YOLOv5-Darknet53 EfficientDet5 COCO AP (%) number of parameters (M) figure 1: speed-accuracy trade-off of accurate models (top) and Size-accuracy curve of lite models on mobile devices (bottom) for YOLOX and other state-of-the-art object detectors\n          \n          \n            \n              Ge\n              Z.\n            \n            \n              Liu\n              S.\n            \n            \n              Wang\n              F.\n            \n            \n              Li\n              Z.\n            \n            \n              Sun\n              J.\n            \n          \n          \n          2025"},{"key":"B14","doi-asserted-by":"publisher","first-page":"1559987","DOI":"10.3389\/fbioe.2025.1559987","article-title":"Automated detection of pinworm parasite eggs using YOLO convolutional block attention module for enhanced microscopic image analysis","volume":"13","author":"Hassan","year":"2025","journal-title":"Front. Bioeng. Biotechnol."},{"key":"B15","doi-asserted-by":"publisher","first-page":"11651","DOI":"10.1038\/s41598-017-11534-0","article-title":"Real-time image processing for microscopy-based label-free imaging flow cytometry in a microfluidic chip","volume":"7","author":"Heo","year":"2017","journal-title":"Sci. Rep."},{"key":"B16","doi-asserted-by":"crossref","DOI":"10.3389\/fbioe.2025.1608194","article-title":"Identification of critical process parameters and quality attributes for bioreactor-based expansion of human MSCs","volume-title":"Frontiers in Bioengineering and Biotechnology","author":"Herbst","year":"2025"},{"key":"B17","doi-asserted-by":"publisher","first-page":"226","DOI":"10.4236\/jcc.2023.117014","article-title":"Research on traffic sign detection based on improved YOLOv8","volume":"11","author":"Huang","year":"2023","journal-title":"J. Comput. Commun."},{"key":"B18","article-title":"International Conference of Harmonisation of Technical requeriments for registration of pharmaceuticals for human use. Ich Q2(R1)","year":"2026"},{"key":"B19","doi-asserted-by":"publisher","first-page":"100876","DOI":"10.1016\/j.xcrp.2022.100876","article-title":"Performance and limitations of deep learning semantic segmentation of multiple defects in transmission electron micrographs","volume":"3","author":"Jacobs","year":"2022","journal-title":"Cell Rep. Phys. Sci."},{"key":"B20","unstructured":"YOLO evolution: a comprehensive benchmark and architectural review of YOLOv12, YOLO11, and their previous versions\n          \n          \n            \n              Jegham\n              N.\n            \n            \n              Koh\n              C. Y.\n            \n            \n              Abdelatti\n              M.\n            \n            \n              Hendawi\n              A.\n            \n          \n          \n          2025"},{"key":"B21","volume-title":"YOLOv11: An Overview of the Key Architectural Enhancements","author":"Khanam","year":"2024"},{"key":"B22","doi-asserted-by":"publisher","first-page":"1520972","DOI":"10.3389\/fonc.2025.1520972","article-title":"Leveraging automated time-lapse microscopy coupled with deep learning to automate colony forming assay","volume":"15","author":"Klett","year":"2025","journal-title":"Front. Oncol."},{"key":"B23","doi-asserted-by":"publisher","first-page":"11307","DOI":"10.1038\/s41598-020-67529-x","article-title":"Object detection based on an adaptive attention mechanism","volume":"10","author":"Li","year":"2020","journal-title":"Sci. Rep."},{"key":"B24","volume-title":"LNCS 8693 - Microsoft COCO: Common Objects in Context","author":"Lin","year":"2014"},{"key":"B25","doi-asserted-by":"publisher","first-page":"2300381","DOI":"10.1002\/biot.202300381","article-title":"Identification of critical process parameters for expansion of clinical grade human Wharton\u2019s jelly-derived mesenchymal stromal cells in stirred-tank bioreactors","volume":"19","author":"L\u00f3pez-Fern\u00e1ndez","year":"2024","journal-title":"Biotechnol. J."},{"key":"B26","volume-title":"SGDR: Stochastic Gradient Descent with Warm Restarts","author":"Loshchilov","year":"2017"},{"key":"B27","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1007\/978-3-031-16852-9_2","article-title":"Benchmarking and boosting transformers for medical image classification","volume":"13542","author":"Ma","year":"2022","journal-title":"Lect. Notes Comput. Sci. Incl. Subser."},{"key":"B28","doi-asserted-by":"publisher","first-page":"2270","DOI":"10.3390\/s25072270","article-title":"YOLO object detection for real-time fabric defect inspection in the textile industry: a review of YOLOv1 to YOLOv11","volume":"25","author":"Mao","year":"2025","journal-title":"Sensors"},{"key":"B29","unstructured":"Deep transfer learning for image classification: a survey\n          \n          \n            \n              Plested\n              J.\n            \n            \n              Phiri\n              M.\n            \n            \n              Gedeon\n              T.\n            \n          \n          \n          2025"},{"key":"B30","volume-title":"Real-Time Flying Object Detection with YOLOv8","author":"Reis","year":"2024"},{"key":"B31","doi-asserted-by":"publisher","first-page":"622","DOI":"10.1136\/ard.56.10.622","article-title":"Comparison of manual and automated cell counts in EDTA preserved synovial fluids. Storage has little influence on the results","volume":"56","author":"Salinas","year":"1997","journal-title":"Ann. Rheum. Dis."},{"key":"B32","doi-asserted-by":"publisher","first-page":"22","DOI":"10.3390\/diagnostics15010022","article-title":"Automated blood cell detection and classification in microscopic images using YOLOv11 and optimized weights","volume":"15","author":"Sazak","year":"2025","journal-title":"Diagnostics"},{"key":"B33","doi-asserted-by":"publisher","first-page":"517","DOI":"10.12688\/f1000research.12533.1","article-title":"Recent advances in bioreactors for cell-based therapies","volume":"7","author":"Stephenson","year":"2018","journal-title":"F1000Research"},{"key":"B34","doi-asserted-by":"publisher","first-page":"1854","DOI":"10.1080\/03036758.2024.2424802","article-title":"YOLOv8E: an efficient YOLOv8 method for instance segmentation of individual tree crowns in Wellington City, New Zealand","volume":"55","author":"Sun","year":"2025","journal-title":"J. R. Soc. N. Z."},{"key":"B35","volume-title":"A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS","author":"Terven","year":"2024"},{"key":"B36","doi-asserted-by":"publisher","first-page":"1922","DOI":"10.1109\/TPAMI.2020.3032166","article-title":"FCOS: a simple and strong anchor-free object detector","volume":"44","author":"Tian","year":"2022","journal-title":"IEEE Trans. Pattern Analysis Mach. Intell."},{"key":"B37","doi-asserted-by":"publisher","first-page":"96","DOI":"10.3390\/bioengineering8070096","article-title":"Bioprocessing of human mesenchymal stem cells: from planar culture to microcarrier-based bioreactors","volume":"8","author":"Tsai","year":"2021","journal-title":"Bioengineering"},{"key":"B38","unstructured":"2025"},{"key":"B39","unstructured":"YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors\n          \n          \n            \n              Wang\n              C.-Y.\n            \n            \n              Bochkovskiy\n              A.\n            \n            \n              Liao\n              H.-Y. M.\n            \n          \n          \n          2022"},{"key":"B40","volume-title":"RT-DETRv3: Real-Time end-to-end Object Detection with Hierarchical Dense Positive Supervision","author":"Wang","year":"2024"},{"key":"B41","doi-asserted-by":"publisher","first-page":"e26586","DOI":"10.1016\/j.heliyon.2024.e26586","article-title":"Improving classification accuracy of fine-tuned CNN models: impact of hyperparameter optimization","volume":"10","author":"Wojciuk","year":"2024","journal-title":"Heliyon"},{"key":"B42","unstructured":"2025"},{"key":"B43","unstructured":"Best practices of successive halving on neural machine translation and large language models\n          \n          \n            \n              Zhang\n              X.\n            \n            \n              Duh\n              K.\n            \n          \n          \n          2024"},{"key":"B44","doi-asserted-by":"publisher","first-page":"29720","DOI":"10.1038\/s41598-025-13960-x","article-title":"Balancing complexity and accuracy for defect detection on filters with an improved RT-DETR","volume":"15","author":"Zhang","year":"2025","journal-title":"Sci. Rep."},{"key":"B45","unstructured":"DETRs beat YOLOs on real-time object detection\n          \n          \n            \n              Zhao\n              Y.\n            \n            \n              Lv\n              W.\n            \n            \n              Xu\n              S.\n            \n            \n              Wei\n              J.\n            \n            \n              Wang\n              G.\n            \n            \n              Dang\n              Q.\n            \n          \n          \n          2024"},{"key":"B46","article-title":"Distance-IoU loss: faster and better learning for bounding box regression","volume-title":"Redmon and Farhadi","author":"Zheng","year":"2016"},{"key":"B47","volume-title":"Random Erasing Data Augmentation","author":"Zhong","year":"2017"}],"container-title":["Frontiers in Bioengineering and Biotechnology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbioe.2026.1849366\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T05:36:44Z","timestamp":1783575404000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbioe.2026.1849366\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,9]]},"references-count":47,"alternative-id":["10.3389\/fbioe.2026.1849366"],"URL":"https:\/\/doi.org\/10.3389\/fbioe.2026.1849366","relation":{},"ISSN":["2296-4185"],"issn-type":[{"value":"2296-4185","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,9]]},"article-number":"1849366"}}