{"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/deepigeos-a-deep-interactive-geodesic","title":"DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation","arxiv_id":"1707.00652","date":"2017-07-03","proceeding":null,"authors":["Guotai Wang","Maria A. Zuluaga","Wenqi Li","Rosalind Pratt","Premal A. Patel","Michael Aertsen","Tom Doel","Anna L. David","Jan Deprest","Sebastien Ourselin","Tom Vercauteren"],"abstract":"Accurate medical image segmentation is essential for diagnosis, surgical\nplanning and many other applications. Convolutional Neural Networks (CNNs) have\nbecome the state-of-the-art automatic segmentation methods. However, fully\nautomatic results may still need to be refined to become accurate and robust\nenough for clinical use. We propose a deep learning-based interactive\nsegmentation method to improve the results obtained by an automatic CNN and to\nreduce user interactions during refinement for higher accuracy. We use one CNN\nto obtain an initial automatic segmentation, on which user interactions are\nadded to indicate mis-segmentations. Another CNN takes as input the user\ninteractions with the initial segmentation and gives a refined result. We\npropose to combine user interactions with CNNs through geodesic distance\ntransforms, and propose a resolution-preserving network that gives a better\ndense prediction. In addition, we integrate user interactions as hard\nconstraints into a back-propagatable Conditional Random Field. We validated the\nproposed framework in the context of 2D placenta segmentation from fetal MRI\nand 3D brain tumor segmentation from FLAIR images. Experimental results show\nour method achieves a large improvement from automatic CNNs, and obtains\ncomparable and even higher accuracy with fewer user interventions and less time\ncompared with traditional interactive methods.","url_abs":"http://arxiv.org/abs/1707.00652v3","url_pdf":"http://arxiv.org/pdf/1707.00652v3.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":"deepigeos-a-deep-interactive-geodesic","repo_url":"https://github.com/taigw/geodesic_distance","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"brain-tumor-segmentation","task_name":"Brain Tumor Segmentation"},{"task_slug":"image-segmentation","task_name":"Image Segmentation"},{"task_slug":"interactive-segmentation","task_name":"Interactive Segmentation"},{"task_slug":"medical-image-segmentation","task_name":"Medical Image Segmentation"},{"task_slug":"placenta-segmentation","task_name":"Placenta Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"tumor-segmentation","task_name":"Tumor Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.00652","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}