dc.creator | Sovatzidi G., Vasilakakis M.D., Iakovidis D.K. | en |
dc.date.accessioned | 2023-01-31T09:59:38Z | |
dc.date.available | 2023-01-31T09:59:38Z | |
dc.date.issued | 2022 | |
dc.identifier | 10.1007/978-3-031-17979-2_8 | |
dc.identifier.isbn | 9783031179785 | |
dc.identifier.issn | 03029743 | |
dc.identifier.uri | http://hdl.handle.net/11615/79235 | |
dc.description.abstract | The detection of abnormalities in the gastrointestinal (GI) tract, including precancerous lesions, is substantially subject to expert knowledge and experience. To address the challenge of automated lesion risk assessment, based on Wireless Capsule Endoscopy (WCE) images, this paper introduces a novel Artificial Intelligence (AI) framework based on Fuzzy Cognitive Maps (FCMs). Specifically, FCMs are fuzzy graph structures used to model knowledge spaces using cause-and-effect relationships, enabling uncertainty-aware reasoning and inference. The novel proposed Interpretable FCM-based Feature Fusion (IF3) framework, includes the following contributions: a) it automatically constructs an FCM based on similarities discovered in training data; b) it enables the fusion of different features extracted using different methods. The proposed framework is generic, domain-independent and it can be integrated into any classifier. To demonstrate its performance, experiments were conducted using real datasets, which include a variety of GI abnormalities, and different feature extractors. The results show that the automatically constructed FCM outperforms state-of-the-art methods, while providing interpretable results, in an easily understandable way. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG. | en |
dc.language.iso | en | en |
dc.source | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | en |
dc.source.uri | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85140440247&doi=10.1007%2f978-3-031-17979-2_8&partnerID=40&md5=4b688a7983e8432f1d92cbefb91e38ec | |
dc.subject | Endoscopy | en |
dc.subject | Fuzzy rules | en |
dc.subject | Risk assessment | en |
dc.subject | Expert experience | en |
dc.subject | Expert knowledge | en |
dc.subject | Features fusions | en |
dc.subject | Fuzzy graph | en |
dc.subject | Gastrointestinal tract | en |
dc.subject | Interpretability | en |
dc.subject | Knowledge and experience | en |
dc.subject | Precancerous lesion | en |
dc.subject | Risks assessments | en |
dc.subject | Wireless capsule endoscopy image | en |
dc.subject | Fuzzy Cognitive Maps | en |
dc.subject | Springer Science and Business Media Deutschland GmbH | en |
dc.title | IF3: An Interpretable Feature Fusion Framework for Lesion Risk Assessment Based on Auto-constructed Fuzzy Cognitive Maps | en |
dc.type | conferenceItem | en |