2026-06-24T10:55:31Zhttps://docta.ucm.es/rest/oai/request
oai:docta.ucm.es:20.500.14352/1237502025-11-11T13:17:51Zcom_20.500.14352_14col_20.500.14352_15
Docta Complutense author Redondo, Alejandro author Ivaylova, Katerina author Bachiller, Margarita author Rincón, Mariano author Cuadra, Jose Manuel author Tamimi, Faleh author López-Cedrún, Jose Luis author Diniz-Freitas, Márcio author Lago-Méndez, Lucía author Rubín-Roger, Guillermo author Torres García Denche, Jesús author Bagan, Leticia author Hernández Vallejo, Gonzalo author López-Pintor Muñoz, Rosa María 2025-09-08T08:00:53Z 2025-09-08T08:00:53Z 2025-07-18 Redondo A, Ivaylova K, Bachiller M, Rincón M, Cuadra JM, Tamimi F, López-Cedrún JL, Diniz-Freitas M, Lago-Méndez L, Rubín-Roger G, Torres J, Bagán L, Hernández G, López-Pintor RM. Multiclass classification of oral mucosal lesions by deep learning from clinical images without performing any restrictions. Biomedical Signal Processing and Control. 2026 En;111(108337) 1746-8094 10.1016/j.bspc.2025.108337 https://hdl.handle.net/20.500.14352/123750 1746-8108 https://doi.org/10.1016/j.bspc.2025.108337 https://www.sciencedirect.com/science/article/pii/S1746809425008481#da1 Oral cancer is a frequently malignant tumor that can be detected during an oral examination. Unfortunately, it is often diagnosed in advanced stages, which leads to low survival rates of about 50% at five years. Due to the low survival rate, it is crucial to develop automated systems that allow the classification of oral lesions according to their severity, aiding in the early diagnosis of oral cancer. This study aims to investigate the effectiveness of using clinical images and deep learning based models to perform a multiclass classification of oral mucosal lesions in color photographs taken without following any acquisition protocol. The classification differentiated four classes: malignant, potentially malignant, benign and healthy. The dataset included a total of 3246 images from 1013 patients, with 40 different categories of oral lesions, including healthy oral mucosa. The images showed different areas of the oral cavity and were captured from different perspectives by diverse dentists and maxillofacial surgeons in the practice. For the classification, different deep learning architectures were applied and compared, from the best known convolutional neural networks (CNN) and skip connection networks (SCN), to more innovative architectures such as visual transformers and a recent hybrid architecture, ConvNeXt v2. The ConvNeXt v2 Tiny architecture, with 85.53% accuracy, 85.02% precision, 85.50% recall, 84.92% F1-score, and 97.40% ROC AUC for an input image size of 354 × 354 pixels, outperformed the other architectures on the same database. The present model improved on previous proposals by considering a greater number of oral lesions and output classes. eng Attribution-NonCommercial-NoDerivatives 4.0 International Multiclass classification of oral mucosal lesions by deep learning from clinical images without performing any restrictions journal article UGFyYSBxdWUgZWwgcmVwb3NpdG9yaW8gaW5zdGl0dWNpb25hbCBkZSBsYSBVbml2ZXJzaWRhZCBkZSBNYWRyaWQgcHVlZGEgYWxtYWNlbmFyIHkgZGlzdHJpYnVpciBlbCBvYmpldG8gZGlnaXRhbCBkZXBvc2l0YWRvLCBlcyBuZWNlc2FyaW8gcXVlIGxhIHBlcnNvbmEgcXVlIGhhZ2EgZWwgZGVww7NzaXRvIGxlYSB5IGFjZXB0ZSBsYXMgY29uZGljaW9uZXMgZXN0YWJsZWNpZGFzIGVuIGVzdGEgbGljZW5jaWE6CuKAoiBBY2VwdGFuZG8gZXN0YSBsaWNlbmNpYSwgdXN0ZWQgY29tbyBhdXRvci9yZXMgeS9vIHByb3BpZXRhcmlvL3MgZGUgbG9zIGRlcmVjaG9zIGRlIHByb3BpZWRhZCBpbnRlbGVjdHVhbCwgbyBsYSBwZXJzb25hIGRlbGVnYWRhIHBhcmEgaGFjZXJsbyAsIG90b3JnYSBhbCByZXBvc2l0b3JpbyBkZSBsYSBVbml2ZXJzaWRhZCBDb21wbHV0ZW5zZSBkZSBNYWRyaWQgZWwgZGVyZWNobyBubyBleGNsdXNpdm8gZGUgcmVwcm9kdWNjacOzbiB5IGRpc3RyaWJ1Y2nDs24gZGUgc3Ugb2JyYSBlbiBjdWFscXVpZXIgZm9ybWF0byB5IG1lZGlvLgrigKIgRWwgZGVwb3NpdGFudGUsIGVuIGNhc28gZGUgdW5hIG9icmEgY29uIG3DoXMgZGUgdW4gYXV0b3IsIGdhcmFudGl6YSBxdWUgbG8gaGFjZSByZXNwb25zYWJsZW1lbnRlIGVuIG5vbWJyZSB5IGNvbiBjb25zZW50aW1pZW50byBkZSBsb3MgZGVtw6FzIGF1dG9yZXMuCuKAoiBBY2VwdGEgcXVlIGVsIHJlcG9zaXRvcmlvIGRlIGxhIFVuaXZlcnNpZGFkIENvbXBsdXRlbnNlIGRlIE1hZHJpZCBwdWVkYSBjb25zZXJ2YXIgbcOhcyBkZSB1bmEgY29waWEgZGUgZXN0ZSBkb2N1bWVudG8geSwgc2luIGFsdGVyYXIgc3UgY29udGVuaWRvLCBjb252ZXJ0aXJsbyBhIGN1YWxxdWllciBmb3JtYXRvIGRlIGZpY2hlcm8sIG1lZGlvIG8gc29wb3J0ZSwgcGFyYSBwcm9ww7NzaXRvcyBkZSBzZWd1cmlkYWQgeSBwcmVzZXJ2YWNpw7NuLgrigKIgRGVjbGFyYSBxdWUgZWwgZG9jdW1lbnRvIGVzIHVuIHRyYWJham8gb3JpZ2luYWwgc3V5bywgbyBxdWUgZXMgbGEgcGVyc29uYSBkZWxlZ2FkYSBwYXJhIHJlYWxpemFyIGVsIGRlcMOzc2l0byB5LCBwb3IgdGFudG8sIGN1ZW50YSBjb24gbGEgcG90ZXN0YWQgcGFyYSBvdG9yZ2FyIGxvcyBkZXJlY2hvcyBjb250ZW5pZG9zIGVuIGVzdGEgbGljZW5jaWEuIFRhbWJpw6luIGRlY2xhcmEgcXVlIHN1IGRvY3VtZW50byBubyBpbmZyaW5nZSBsbyBlc3RhYmxlY2lkbyBlbiBsYSB2aWdlbnRlIGxlZ2lzbGFjacOzbiBkZSBwcm9waWVkYWQgaW50ZWxlY3R1YWwuCuKAoiBTaSBlbCBkb2N1bWVudG8gY29udGllbmUgbWF0ZXJpYWxlcyBkZSBsb3MgY3VhbGVzIG5vIHRpZW5lIGxvcyBkZXJlY2hvcyBkZSBhdXRvciwgZGVjbGFyYSBxdWUgaGEgb2J0ZW5pZG8gZWwgcGVybWlzbyBzaW4gcmVzdHJpY2Npw7NuIGRlbCBwcm9waWV0YXJpbyBwYXJhIG90b3JnYXIgYSBsYSBVbml2ZXJzaWRhZCBDb21wbHV0ZW5zZSBkZSBNYWRyaWQgbG9zIGRlcmVjaG9zIHJlcXVlcmlkb3MgcG9yIGVzdGEgbGljZW5jaWEsIHkgcXVlIGVzZSBtYXRlcmlhbCwgY3V5b3MgZGVyZWNob3Mgc29uIGRlIHRlcmNlcm9zLCBlc3TDoSBjbGFyYW1lbnRlIGlkZW50aWZpY2FkbyB5IHJlY29ub2NpZG8gZGVudHJvIGRlbCBjb250ZW5pZG8gZGVsIGRvY3VtZW50by4K4oCiIFNpIGVsIGRvY3VtZW50byBzZSBiYXNhIGVuIHVuYSBvYnJhIHF1ZSBoYSBzaWRvIHBhdHJvY2luYWRhIG8gYXBveWFkYSBwb3IgdW5hIGFnZW5jaWEgdSBvcmdhbml6YWNpw7NuIGRpZmVyZW50ZSBkZSBsYSBVbml2ZXJzaWRhZCBDb21wbHV0ZW5zZSBkZSBNYWRyaWQsIHVzdGVkIGRlY2xhcmEgcXVlIGhhIGN1bXBsaWRvIGNvbiBsYXMgb2JsaWdhY2lvbmVzIHJlcXVlcmlkYXMgcG9yIGVzdGUgY29udHJhdG8gbyBhY3VlcmRvLgrigKIgTGEgVW5pdmVyc2lkYWQgQ29tcGx1dGVuc2UgZGUgTWFkcmlkIGlkZW50aWZpY2Fyw6EgY2xhcmFtZW50ZSBzdS9zIG5vbWJyZS9zIGNvbW8gZWwvbG9zIGF1dG9yL2VzIG8gcHJvcGlldGFyaW8vcyBkZSBsb3MgZGVyZWNob3MgZGVsIGRvY3VtZW50bywgeSBubyBoYXLDoSBuaW5ndW5hIGFsdGVyYWNpw7NuIGRlIHN1IGRvY3VtZW50byBkaWZlcmVudGUgYSBsYXMgcGVybWl0aWRhcyBlbiBlc3RhIGxpY2VuY2lhLgrigKIgVHJhYmFqb3MgZGVwb3NpdGFkb3MgcG9yIHN1IHByb3BpbyBhdXRvcjogYWwgYXV0by1hcmNoaXZhciBlc3RhIGNvbGVjY2nDs24gZGUgZmljaGVyb3MgeSBtZXRhZGF0b3MgYmlibGlvZ3LDoWZpY29zIGFzb2NpYWRvcywgb3RvcmdvIGFsIHJlcG9zaXRvcmlvIGluc3RpdHVjaW9uYWwgZGUgbGEgVW5pdmVyc2lkYWQgQ29tcGx1dGVuc2UgZGUgTWFkcmlkIGVsIGRlcmVjaG8gcGFyYSBhbG1hY2VuYXJsb3MgeSBkaXN0cmlidWlybG9zLiBEZWNsYXJvIHF1ZSBlc3RlIG1hdGVyaWFsIGVzIGRlIG1pIHByb3BpZWRhZCBpbnRlbGVjdHVhbCB5IHF1ZSBlbCByZXBvc2l0b3JpbyBpbnN0aXR1dGljaW9uYWwgZGUgbGEgVW5pdmVyc2lkYWQgQ29tcGx1dGVuc2UgZGUgTWFkcmlkLCBubyBhc3VtZSByZXNwb25zYWJpbGlkYWQgYWxndW5hIGVuIGNhc28gZGUgcXVlIHNlIHByb2R1emNhIHVuYSB2aW9sYWNpw7NuIGRlIGRlcmVjaG9zIGRlIHByb3BpZWRhZCBhbCBkaXN0cmlidWlyIGVzdG9zIGZpY2hlcm9zIG8gbWV0YWRhdG9zLgrigKIgVHJhYmFqb3MgZGVwb3NpdGFkb3MgcG9yIG90cm9zIHF1ZSBubyBzZWFuIHN1IGF1dG9yOiBkZWNsYXJvIHF1ZSBsb3MgbWV0YWRhdG9zIGJpYmxpb2dyw6FmaWNvcyBxdWUgZXN0b3kgYXJjaGl2YW5kbyBlbiBlbCByZXBvc2l0b3JpbyBpbnN0aXR1Y2lvbmFsIGRlIGxhIFVuaXZlcnNpZGFkIENvbXBsdXRlbnNlIGRlIE1hZHJpZCwgc29uIGRlIGRvbWluaW8gcMO6YmxpY28uIFNpIG5vIGZ1ZXNlIGVsIGNhc28sIGFjZXB0byBwbGVuYSByZXNwb25zYWJpbGlkYWQgcG9yIGN1YWxxdWllciBpbmZyYWNjacOzbiBkZSBkZXJlY2hvcyBkZSBhdXRvciBxdWUgY29ubGxldmUgbGEgZGlzdHJpYnVjacOzbiBkZSBlc3RvcyBmaWNoZXJvcyBvIG1ldGFkYXRvcy4K4oCiIFBvciB0YW50bywgdG9kb3MgbG9zIG1ldGFkYXRvcyBwdWJsaWNhZG9zIGVuIGVsIHJlcG9zaXRvcmlvIGluc3RpdHVjaW9uYWwgZGUgbGEgVW5pdmVyc2lkYWQgQ29tcGx1dGVuc2UgZGUgTWFkcmlkLCBlc3TDoW4gZGlzcG9uaWJsZXMgY29uIGxhIGxpY2VuY2lhIENyZWF0aXZlIENvbW1vbnMgQ0MwIDEuMCBVbml2ZXJzYWwgT2ZyZWNpbWllbnRvIGFsIERvbWluaW8gUMO6YmxpY28gKENDMCAxLjApLiBFc3RhIGxpY2VuY2lhIHBlcm1pdGUgY29waWFyLCBtb2RpZmljYXIsIGRpc3RyaWJ1aXIsIHJldXRpbGl6YXIgeSBjb21wYXJ0aXIgcMO6YmxpY2EgeSBsaWJyZW1lbnRlIGxvcyBtZXRhZGF0b3MgKGluY2x1eWVuZG8gc3UgdXNvIGNvbWVyY2lhbCksIHNpbiB0ZW5lciBxdWUgcGVkaXIgcGVybWlzby4gU2kgdGVyY2VyYXMgcGFydGVzIHVzYW4gZXN0b3MgbWV0YWRhdG9zLCBkZWJlbiBoYWNlciB1biByZWNvbm9jaW1pZW50byB5IHVuYSBhdHJpYnVjacOzbiBhIGxhcyBmdWVudGVzIGRlIGxvcyBtaXNtb3MuIExhIHJldXRpbGl6YWNpw7NuIGRlIGVzdG9zIG1ldGFkYXRvcyB0YW1iacOpbiBlcyBsaWJyZSwgdGFsIHkgY29tbyBlc3RhYmxlY2UgbGEgbGljZW5jaWEgZGUgdXNvIENDMC4K URL https://docta.ucm.es/bitstreams/24ed2038-4bdc-4552-9940-f23eff7b1b7b/download File MD5 69cb20d573d556896ba12acd9a885606 1562395 application/pdf Biomedical Signal Processing and Control. 2026 En;111(108337).pdf