{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,2]],"date-time":"2026-06-02T08:58:03Z","timestamp":1780390683493,"version":"3.54.1"},"reference-count":44,"publisher":"Wiley","issue":"7","license":[{"start":{"date-parts":[[2025,5,12]],"date-time":"2025-05-12T00:00:00Z","timestamp":1747008000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["1R01CA256890"],"award-info":[{"award-number":["1R01CA256890"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["R01CA275772"],"award-info":[{"award-number":["R01CA275772"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2025,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Deep learning networks map input data to output predictions by fitting network parameters using training data. However, applying a trained network to new, unseen inference data resembles an interpolation process, which may lead to inaccurate predictions if the training and inference data distributions differ significantly.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>This study aims to generally improve the prediction accuracy of deep learning networks on the inference case by bridging the gap between training and inference data.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We propose an inference\u2010specific learning strategy to enhance the network learning process without modifying the network structure. By aligning training data to closely match the specific inference data, we generate an inference\u2010specific training dataset, enhancing the network optimization around the inference data point for more accurate predictions. Taking medical image auto\u2010segmentation as an example, we develop an inference\u2010specific auto\u2010segmentation framework consisting of initial segmentation learning, inference\u2010specific training data deformation, and inference\u2010specific segmentation refinement. The framework is evaluated on public abdominal, head\u2010neck, and pancreas CT datasets comprising 30, 42, and 210 cases, respectively, for medical image segmentation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Experimental results show that our method improves the organ\u2010averaged mean Dice by 6.2% (\n                      <jats:italic>p<\/jats:italic>\n                      \u2010value\u00a0=\u00a00.001), 1.5% (\n                      <jats:italic>p<\/jats:italic>\n                      \u2010value\u00a0=\u00a00.003), and 3.7% (\n                      <jats:italic>p<\/jats:italic>\n                      \u2010value\u00a0&lt;\u00a00.001) on the three datasets, respectively, with a more notable increase for difficult\u2010to\u2010segment organs (such as a 21.7% increase for the gallbladder [\n                      <jats:italic>p<\/jats:italic>\n                      \u2010value\u00a0=\u00a00.004]). By incorporating organ mask\u2010based weak supervision into the training data alignment learning, the inference\u2010specific auto\u2010segmentation accuracy is generally improved compared with the image intensity\u2010based alignment. Besides, a moving\u2010averaged calculation of the inference organ mask during the learning process strengthens both the robustness and accuracy of the final inference segmentation.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>By leveraging inference data during training, the proposed inference\u2010specific learning strategy consistently improves auto\u2010segmentation accuracy and holds the potential to be broadly applied for enhanced deep learning decision\u2010making.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.17883","type":"journal-article","created":{"date-parts":[[2025,5,27]],"date-time":"2025-05-27T08:48:32Z","timestamp":1748335712000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Inference\u2010specific learning for improved medical image segmentation"],"prefix":"10.1002","volume":"52","author":[{"given":"Yizheng","family":"Chen","sequence":"first","affiliation":[{"name":"Department of Radiation Oncology Stanford University Stanford California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sheng","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Stanford University Stanford California USA"},{"name":"Department of Biomedical Data Science Stanford University Stanford California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mingjie","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Stanford University Stanford California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bin","family":"Han","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Stanford University Stanford California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lei","family":"Xing","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Stanford University Stanford California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,5,12]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539"},{"key":"e_1_2_8_3_1","doi-asserted-by":"crossref","unstructured":"KirillovA MintunE RaviN et\u00a0al.Segment anything. In:Proceedings of the IEEE\/CVF InternationalConference on Computer Vision 2023(pp.4015\u20104026); 2023.","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/S2589-7500(24)00114-6"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-021-04223-6"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/TTS.2023.3234203"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15352"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2020.101907"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3358295"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2022.11.004"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.16056"},{"key":"e_1_2_8_12_1","doi-asserted-by":"crossref","unstructured":"RonnebergerO FischerP BroxT.U\u2010net: convolutional networks for biomedical image segmentation. In:Proceedings ofMedical Image Computing and Computer\u2010Assisted Intervention\u2013MICCAI2015: 18th International Conference Munich Germany October 5\u20139 2015 Part III (pp.234\u2010241).Springer International Publishing; 2015.","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-33178-z"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-019-0099-z"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.12045"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-34257-x"},{"key":"e_1_2_8_17_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-26216-9"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-023-00682-w"},{"key":"e_1_2_8_19_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-44824-z"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2024.103182"},{"key":"e_1_2_8_21_1","doi-asserted-by":"crossref","unstructured":"ZhaoA BalakrishnanG DurandF GuttagJV DalcaAV.Data augmentation using learned transformations for one\u2010shot medical image segmentation. In:Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(pp.8543\u20108553);2019.","DOI":"10.1109\/CVPR.2019.00874"},{"key":"e_1_2_8_22_1","doi-asserted-by":"crossref","unstructured":"WangS CaoS WeiD et\u00a0al.LT\u2010Net: label transfer by learning reversible voxel\u2010wise correspondence for one\u2010shot medical image segmentation. In:Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition(pp.9162\u20109171); 2020.","DOI":"10.1109\/CVPR42600.2020.00918"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1111\/1754-9485.13261"},{"key":"e_1_2_8_24_1","doi-asserted-by":"crossref","unstructured":"MaW ChenC ZhengS QinJ ZhangH DouQ.Test\u2010time adaptation with calibration of medical image classification nets for label distribution shift. In:International Conference on Medical Image Computing and Computer\u2010Assisted Intervention(pp.313\u2010323).Springer Nature Switzerland; 2022.","DOI":"10.1007\/978-3-031-16437-8_30"},{"key":"e_1_2_8_25_1","doi-asserted-by":"crossref","unstructured":"WangX GirshickR GuptaA HeK.Non\u2010local neural networks. In:Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(pp.7794\u20107803); 2018.","DOI":"10.1109\/CVPR.2018.00813"},{"key":"e_1_2_8_26_1","doi-asserted-by":"crossref","unstructured":"WangZ ZouN ShenD JiS.Non\u2010local U\u2010Nets for biomedical image segmentation. In:Proceedings of the AAAI Conference on Artificial Intelligence. Vol.34. 4th ed.;2020:6315\u20106322.","DOI":"10.1609\/aaai.v34i04.6100"},{"key":"e_1_2_8_27_1","doi-asserted-by":"crossref","unstructured":"CaoH WangY ChenJ et\u00a0al.Swin\u2010Unet: uNet\u2010like pure transformer for medical image segmentation. In:European Conference on Computer Vision(pp.205\u2010218).Springer Nature Switzerland;2022.","DOI":"10.1007\/978-3-031-25066-8_9"},{"key":"e_1_2_8_28_1","unstructured":"ChenJ LuY YuQ et\u00a0al.Transunet: Transformers Make Strong Encoders for Medical Image Segmentation.arXiv preprint arXiv:2102.04306;2021."},{"key":"e_1_2_8_29_1","unstructured":"OktayO SchlemperJ FolgocLL et\u00a0al.Attention u\u2010net: Learning where to Look for the Pancreas.arXiv preprint arXiv:1804.03999;2018."},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cag.2020.05.003"},{"key":"e_1_2_8_31_1","doi-asserted-by":"crossref","unstructured":"LiC TanY ChenW et\u00a0al.Attention UNet++: a nested attention\u2010aware u\u2010net for liver ct image segmentation. In:2020 IEEE International Conference on Image Processing (ICIP)(pp.345\u2010349).IEEE;2020.","DOI":"10.1109\/ICIP40778.2020.9190761"},{"key":"e_1_2_8_32_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278-024-00981-7"},{"key":"e_1_2_8_33_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278-019-00227-x"},{"key":"e_1_2_8_34_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3362879"},{"key":"e_1_2_8_35_1","doi-asserted-by":"crossref","unstructured":"MilletariF NavabN AhmadiSA.V\u2010net: fully convolutional neural networks for volumetric medical image segmentation. In:2016 Fourth International Conference on 3D Vision (3DV)(pp.565\u2010571).IEEE;2016.","DOI":"10.1109\/3DV.2016.79"},{"key":"e_1_2_8_36_1","doi-asserted-by":"crossref","unstructured":"KodymO \u0160pan\u011blM HeroutA.Segmentation of head and neck organs at risk using cnn with batch dice loss. In:Proceedings of the Pattern Recognition: 40th German Conference GCPR; October 9\u201312 2018 Stuttgart Germany(pp.105\u2010114).Springer International Publishing.","DOI":"10.1007\/978-3-030-12939-2_8"},{"key":"e_1_2_8_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2021.102035"},{"key":"e_1_2_8_38_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/ab79c4"},{"key":"e_1_2_8_39_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2016.2610583"},{"key":"e_1_2_8_40_1","unstructured":"KingmaDP BaJ.Adam: A Method for Stochastic Optimization.arXiv preprint arXiv:1412.6980;2014."},{"key":"e_1_2_8_41_1","unstructured":"LandmanB XuZ IgelsiasJ StynerM LangerakT KleinA.Miccai multi\u2010atlas labeling beyond the cranial vault\u2013workshop and challenge.In: Proc. ofMICCAI Multi\u2010Atlas Labeling Beyond Cranial Vault\u2014Workshop Challenge(Vol.5 p. 12);2015."},{"key":"e_1_2_8_42_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.12197"},{"key":"e_1_2_8_43_1","unstructured":"SimpsonAL AntonelliM BakasS et\u00a0al.A Large Annotated Medical Image Dataset for the Development and Evaluation of Segmentation Algorithms;2019.arXiv preprint arXiv:1902.09063."},{"key":"e_1_2_8_44_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2020.3025308"},{"key":"e_1_2_8_45_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13825"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17883","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,8]],"date-time":"2025-12-08T10:32:46Z","timestamp":1765189966000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17883"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,12]]},"references-count":44,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2025,7]]}},"alternative-id":["10.1002\/mp.17883"],"URL":"https:\/\/doi.org\/10.1002\/mp.17883","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,5,12]]},"assertion":[{"value":"2024-11-20","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-03-25","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-05-12","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e17883"}}