{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T04:13:44Z","timestamp":1775016824985,"version":"3.50.1"},"reference-count":28,"publisher":"Springer Science and Business Media LLC","issue":"7","license":[{"start":{"date-parts":[[2023,5,5]],"date-time":"2023-05-05T00:00:00Z","timestamp":1683244800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,5,5]],"date-time":"2023-05-05T00:00:00Z","timestamp":1683244800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"FWF","award":["project DOC 110"],"award-info":[{"award-number":["project DOC 110"]}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Int J CARS"],"abstract":"<jats:title>Abstract<\/jats:title><jats:sec>\n                <jats:title>Purpose<\/jats:title>\n                <jats:p>Automatic liver segmentation is a key component for performing computer-assisted hepatic procedures. The task is challenging due to the high variability in organ appearance, numerous imaging modalities, and limited availability of labels. Moreover, strong generalization performance is required in real-world scenarios. However, existing supervised methods cannot be applied to data not seen during training (i.e. in the wild) because they generalize poorly.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Methods<\/jats:title>\n                <jats:p>We propose to distill knowledge from a powerful model with our novel contrastive distillation scheme. We use a pre-trained large neural network to train our smaller model. A key novelty is to map neighboring slices close together in the latent representation, while mapping distant slices far away. Then, we use ground-truth labels to learn a U-Net style upsampling path and recover the segmentation map.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Results<\/jats:title>\n                <jats:p>The pipeline is proven to be robust enough to perform state-of-the-art inference on target unseen domains. We carried out an extensive experimental validation using six common abdominal datasets, covering multiple modalities, as well as 18 patient datasets from the Innsbruck University Hospital. A sub-second inference time and a data-efficient training pipeline make it possible to scale our method to real-world conditions.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Conclusion<\/jats:title>\n                <jats:p>We propose a novel contrastive distillation scheme for automatic liver segmentation. A limited set of assumptions and superior performance to state-of-the-art techniques make our method a candidate for application to real-world scenarios.<\/jats:p>\n              <\/jats:sec>","DOI":"10.1007\/s11548-023-02912-3","type":"journal-article","created":{"date-parts":[[2023,5,5]],"date-time":"2023-05-05T07:02:09Z","timestamp":1683270129000},"page":"1143-1149","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":5,"title":["Towards liver segmentation in the wild via contrastive distillation"],"prefix":"10.1007","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4600-8000","authenticated-orcid":false,"given":"Stefano","family":"Fogarollo","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0021-5792","authenticated-orcid":false,"given":"Reto","family":"Bale","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7564-6234","authenticated-orcid":false,"given":"Matthias","family":"Harders","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2023,5,5]]},"reference":[{"key":"2912_CR1","doi-asserted-by":"crossref","unstructured":"Conze P-H, et al. (2021)Abdominal multi-organ segmentation with cascaded convolutional and adversarial deep networks. In: AIME","DOI":"10.1016\/j.artmed.2021.102109"},{"key":"2912_CR2","doi-asserted-by":"crossref","unstructured":"Valanarasu JMJ, et al.: (2022) KiU-Net: overcomplete convolutional architectures for biomedical image and volumetric segmentation. IEEE Trans Med Imaging 41","DOI":"10.1109\/TMI.2021.3130469"},{"key":"2912_CR3","doi-asserted-by":"crossref","unstructured":"Tang Y, et al. (2022) Self-supervised pre-training of swin transformers for 3D medical image analysis. In: CVPR","DOI":"10.1109\/CVPR52688.2022.02007"},{"key":"2912_CR4","doi-asserted-by":"crossref","unstructured":"Geirhos R et al. (2020) Shortcut learning in deep neural networks. Nat Mach Intell 11","DOI":"10.1038\/s42256-020-00257-z"},{"key":"2912_CR5","doi-asserted-by":"crossref","unstructured":"Tajbakhsh N et al (2020) Embracing imperfect datasets: a review of deep learning solutions for medical image segmentation. Med Imag Anal 63","DOI":"10.1016\/j.media.2020.101693"},{"key":"2912_CR6","doi-asserted-by":"crossref","unstructured":"Wen Y et al (2021) Rethinking pre-training on medical imaging. J Vis Commun Image Represent 78","DOI":"10.1016\/j.jvcir.2021.103145"},{"key":"2912_CR7","doi-asserted-by":"crossref","unstructured":"Deng J, et al. (2009) Imagenet: a large-scale hierarchical image database. In: CVPR","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"2912_CR8","doi-asserted-by":"crossref","unstructured":"Bilic P, et al. (2023) The liver tumor segmentation benchmark (LiTS). Med Imag Anal 84","DOI":"10.1016\/j.media.2022.102680"},{"key":"2912_CR9","doi-asserted-by":"crossref","unstructured":"Moghbel M, et al. (2018) Review of liver segmentation and computer assisted detection\/diagnosis methods in computed tomography. Artif Intell Rev 50","DOI":"10.1007\/s10462-017-9550-x"},{"key":"2912_CR10","doi-asserted-by":"crossref","unstructured":"Tsai D-Y, et al. (1994) Neural-network-based boundary detection of liver structure in ct images for 3-d visualization. In: IEEE Intl. Conf. on Neural Networks, vol 6","DOI":"10.1109\/ICNN.1994.374895"},{"key":"2912_CR11","unstructured":"Dou Q, et al. (2019) Domain generalization via model-agnostic learning of semantic features. In: NIPS, vol 32"},{"key":"2912_CR12","doi-asserted-by":"crossref","unstructured":"Li C, et al. (2022) Domain generalization on medical imaging classification using episodic training with task augmentation. Comput Biol Med 141","DOI":"10.1016\/j.compbiomed.2021.105144"},{"key":"2912_CR13","doi-asserted-by":"crossref","unstructured":"Zhou Z, et al. (2022) Generalizable cross-modality medical image segmentation via style augmentation and dual normalization. In: CVPR","DOI":"10.1109\/CVPR52688.2022.02019"},{"key":"2912_CR14","doi-asserted-by":"crossref","unstructured":"Zeng D, et al. (2021) Positional contrastive learning for volumetric medical image segmentation. In: MICCAI","DOI":"10.1007\/978-3-030-87196-3_21"},{"key":"2912_CR15","unstructured":"Hamilton M, et al. (2022) Unsupervised semantic segmentation by distilling feature correspondences. In: ICLR"},{"key":"2912_CR16","unstructured":"Truong T, et al. (2021) How transferable are self-supervised features in medical image classification tasks? In: Machine Learning for Health"},{"key":"2912_CR17","doi-asserted-by":"crossref","unstructured":"Caron M, et al. (2021) Emerging properties in self-supervised vision transformers. In: ICCV","DOI":"10.1109\/ICCV48922.2021.00951"},{"key":"2912_CR18","doi-asserted-by":"crossref","unstructured":"Ronneberger O, et al. (2015) U-net: convolutional networks for biomedical image segmentation. In: MICCAI","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"2912_CR19","doi-asserted-by":"crossref","unstructured":"Lin T-Y, Goyal P, Girshick R, He K, Doll\u00e1r P (2017) Focal loss for dense object detection. In: ICCV","DOI":"10.1109\/ICCV.2017.324"},{"key":"2912_CR20","doi-asserted-by":"crossref","unstructured":"Yu J, et al. (2016) UnitBox: an advanced object detection network. In: ACM Intl. Conf. on Multimedia","DOI":"10.1145\/2964284.2967274"},{"key":"2912_CR21","doi-asserted-by":"crossref","unstructured":"Yang Y, et al. (2020) FDA: Fourier domain adaptation for semantic segmentation. In: CVPR","DOI":"10.1109\/CVPR42600.2020.00414"},{"key":"2912_CR22","unstructured":"Landman B, et al. (2015) Miccai multi-atlas labeling beyond the cranial vault\u2014workshop and challenge. In: MICCAI"},{"key":"2912_CR23","doi-asserted-by":"crossref","unstructured":"Kavur A, et al. (2021) CHAOS challenge combined (CT-MR) healthy abdominal organ segmentation. Med Imag Anal 69","DOI":"10.1016\/j.media.2020.101950"},{"key":"2912_CR24","unstructured":"Soler L et al (2010) 3d image reconstruction for comparison of algorithm database: a patient-specific anatomical and medical image database. Tech. Rep, IRCAD, Strasbourg, France"},{"key":"2912_CR25","doi-asserted-by":"crossref","unstructured":"Ma J, et al. (2022) Abdomenct-1k: is abdominal organ segmentation a solved problem? IEEE Trans Pattern Anal Mach Intell 44(10)","DOI":"10.1109\/TPAMI.2021.3100536"},{"key":"2912_CR26","unstructured":"Liu L, et al. (2020) On the variance of the adaptive learning rate and beyond. In: ICLR"},{"key":"2912_CR27","doi-asserted-by":"crossref","unstructured":"P\u00e9rez-Garc\u00eda F, et al. (2021) Torchio: a python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning. Comput Methods Program Biomed 208","DOI":"10.1016\/j.cmpb.2021.106236"},{"key":"2912_CR28","doi-asserted-by":"crossref","unstructured":"Bale R, et al. (2011) Stereotactic radiofrequency ablation. CardioVasc Interv Radiol 34(4)","DOI":"10.1007\/s00270-010-9966-z"}],"container-title":["International Journal of Computer Assisted Radiology and Surgery"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s11548-023-02912-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s11548-023-02912-3\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s11548-023-02912-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,7,8]],"date-time":"2023-07-08T16:08:08Z","timestamp":1688832488000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s11548-023-02912-3"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,5,5]]},"references-count":28,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2023,7]]}},"alternative-id":["2912"],"URL":"https:\/\/doi.org\/10.1007\/s11548-023-02912-3","relation":{},"ISSN":["1861-6429"],"issn-type":[{"value":"1861-6429","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,5,5]]},"assertion":[{"value":"9 March 2023","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 April 2023","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"5 May 2023","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors have no conflict of interest.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}}]}}