{"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/capsules-for-object-segmentation","title":"Capsules for Object Segmentation","arxiv_id":"1804.04241","date":"2018-04-11","proceeding":null,"authors":["Rodney LaLonde","Ulas Bagci"],"abstract":"Convolutional neural networks (CNNs) have shown remarkable results over the\nlast several years for a wide range of computer vision tasks. A new\narchitecture recently introduced by Sabour et al., referred to as a capsule\nnetworks with dynamic routing, has shown great initial results for digit\nrecognition and small image classification. The success of capsule networks\nlies in their ability to preserve more information about the input by replacing\nmax-pooling layers with convolutional strides and dynamic routing, allowing for\npreservation of part-whole relationships in the data. This preservation of the\ninput is demonstrated by reconstructing the input from the output capsule\nvectors. Our work expands the use of capsule networks to the task of object\nsegmentation for the first time in the literature. We extend the idea of\nconvolutional capsules with locally-connected routing and propose the concept\nof deconvolutional capsules. Further, we extend the masked reconstruction to\nreconstruct the positive input class. The proposed\nconvolutional-deconvolutional capsule network, called SegCaps, shows strong\nresults for the task of object segmentation with substantial decrease in\nparameter space. As an example application, we applied the proposed SegCaps to\nsegment pathological lungs from low dose CT scans and compared its accuracy and\nefficiency with other U-Net-based architectures. SegCaps is able to handle\nlarge image sizes (512 x 512) as opposed to baseline capsules (typically less\nthan 32 x 32). The proposed SegCaps reduced the number of parameters of U-Net\narchitecture by 95.4% while still providing a better segmentation accuracy.","url_abs":"http://arxiv.org/abs/1804.04241v1","url_pdf":"http://arxiv.org/pdf/1804.04241v1.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":"capsules-for-object-segmentation","repo_url":"https://github.com/Cheng-Lin-Li/SegCaps","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/Ian-Liao/SegCaps","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/Oushesh/CapsulesforSegmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/Oushesh/SegCaps-TF_Version","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/StuvX/BayesCapsNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/kevins99/SegCaps-Keras","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"capsules-for-object-segmentation","repo_url":"https://github.com/lalonderodney/SegCaps","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.04241","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}