{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/anatomically-constrained-neural-networks-acnn","title":"Anatomically Constrained Neural Networks (ACNN): Application to Cardiac Image Enhancement and Segmentation","arxiv_id":"1705.08302","date":"2017-05-22","proceeding":null,"authors":["Ozan Oktay","Enzo Ferrante","Konstantinos Kamnitsas","Mattias Heinrich","Wenjia Bai","Jose Caballero","Stuart Cook","Antonio de Marvao","Timothy Dawes","Declan O'Regan","Bernhard Kainz","Ben Glocker","Daniel Rueckert"],"abstract":"Incorporation of prior knowledge about organ shape and location is key to\nimprove performance of image analysis approaches. In particular, priors can be\nuseful in cases where images are corrupted and contain artefacts due to\nlimitations in image acquisition. The highly constrained nature of anatomical\nobjects can be well captured with learning based techniques. However, in most\nrecent and promising techniques such as CNN based segmentation it is not\nobvious how to incorporate such prior knowledge. State-of-the-art methods\noperate as pixel-wise classifiers where the training objectives do not\nincorporate the structure and inter-dependencies of the output. To overcome\nthis limitation, we propose a generic training strategy that incorporates\nanatomical prior knowledge into CNNs through a new regularisation model, which\nis trained end-to-end. The new framework encourages models to follow the global\nanatomical properties of the underlying anatomy (e.g. shape, label structure)\nvia learned non-linear representations of the shape. We show that the proposed\napproach can be easily adapted to different analysis tasks (e.g. image\nenhancement, segmentation) and improve the prediction accuracy of the\nstate-of-the-art models. The applicability of our approach is shown on\nmulti-modal cardiac datasets and public benchmarks. Additionally, we\ndemonstrate how the learned deep models of 3D shapes can be interpreted and\nused as biomarkers for classification of cardiac pathologies.","url_abs":"http://arxiv.org/abs/1705.08302v4","url_pdf":"http://arxiv.org/pdf/1705.08302v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"anatomically-constrained-neural-networks-acnn","repo_url":"https://gitlab.com/matzkin/deep-brain-extractor","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"anatomy","task_name":"Anatomy"},{"task_slug":"image-enhancement","task_name":"Image Enhancement"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1705.08302","atlas_url":"https://app.syntology.ai/?focus=1705.08302","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}