Next Article in Journal
Mine Water Hazard Video Recognition Based on Residual Preprocessing and Temporal–Spatial Descriptors
Previous Article in Journal
MOOC Learner Profiling Using Relation-Aware Heterogeneous Graph Neural Networks
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 16
Issue 1
10.3390/app16010264
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
This is an early access version, the complete PDF, HTML, and XML versions will be available soon.
Article

Context-Aware Caries Segmentation in Periapical Radiographs Using a Hybrid Multi-Task Learning Framework with Partial Annotations

by
Jesús Antonio Nava-Pintor
1,
Héctor A. Guerrero-Osuna
1,*,
Fabián García-Vázquez
1,
Luis F. Luque-Vega
2,
Teodoro Ibarra-Pérez
3,
Salvador Ibarra-Delgado
4,
Víktor I. Rodríguez-Abdalá
4,
Remberto Sandoval-Arechiga
4 and
José Ricardo Gómez-Rodríguez
4
1
Posgrado en Ingeniería y Tecnología Aplicada, Unidad Académica de Ingeniería Eléctrica, Universidad Autónoma de Zacatecas, Zacatecas 98000, Zacatecas, Mexico
2
Department of Technological and Industrial Processes, ITESO, Tlaquepaque 45604, Jalisco, Mexico
3
Unidad Profesional Interdisciplinaria de Ingeniería Campus Zacatecas (UPIIZ), Instituto Politécnico Nacional, Zacatecas 98160, Zacatecas, Mexico
4
Unidad Académica de Ingeniería Eléctrica, Universidad Autónoma de Zacatecas, Zacatecas 98000, Zacatecas, Mexico
*
Author to whom correspondence should be addressed.
Appl. Sci. 2026, 16(1), 264; https://doi.org/10.3390/app16010264 (registering DOI)
Submission received: 4 December 2025 / Revised: 15 December 2025 / Accepted: 22 December 2025 / Published: 26 December 2025
(This article belongs to the Special Issue New Trends in Image Classification and Pattern Recognition)
Versions Notes

Abstract

Dental caries remains a prevalent diagnostic challenge, particularly in periapical radiographs, where geometric distortions and anatomical overlap complicate interpretation. Although deep learning has advanced dental image analysis, most segmentation models depend on fully annotated datasets and rarely exploit anatomical context. This study proposes a hybrid Multi-Task Learning (MTL) framework that jointly performs anatomical segmentation and caries detection in scenarios with partial and asymmetric annotations. The method integrates a U-Net++ dual-head architecture with a shared EfficientNet-B4 encoder, supplemented by pseudo-label generation and selective loss masking to handle incomplete ground truth. We hypothesize that learning healthy dental structures provides a contextual scaffold that enhances the identification of pathology. Experiments on an independent test set validate this hypothesis: the proposed MTL model achieved an DSC of 0.6706 and an IoU of 0.5044, outperforming a specialized single-task baseline. Most notably, sensitivity improved by 7.47%, reducing false-negative pixels by 19.9%. These findings demonstrate that context-aware supervision substantially improves detection robustness on complex periapical radiographs, even when full-pixel-level annotations are unavailable. The proposed framework offers a scalable pathway for developing clinically oriented diagnostic tools in real-world settings where annotation completeness is limited.
Keywords: artificial intelligence; semantic segmentation; deep learning; multi-task learning; periapical radiographs; caries segmentation artificial intelligence; semantic segmentation; deep learning; multi-task learning; periapical radiographs; caries segmentation

Share and Cite

MDPI and ACS Style

Nava-Pintor, J.A.; Guerrero-Osuna, H.A.; García-Vázquez, F.; Luque-Vega, L.F.; Ibarra-Pérez, T.; Ibarra-Delgado, S.; Rodríguez-Abdalá, V.I.; Sandoval-Arechiga, R.; Gómez-Rodríguez, J.R. Context-Aware Caries Segmentation in Periapical Radiographs Using a Hybrid Multi-Task Learning Framework with Partial Annotations. Appl. Sci. 2026, 16, 264. https://doi.org/10.3390/app16010264

AMA Style

Nava-Pintor JA, Guerrero-Osuna HA, García-Vázquez F, Luque-Vega LF, Ibarra-Pérez T, Ibarra-Delgado S, Rodríguez-Abdalá VI, Sandoval-Arechiga R, Gómez-Rodríguez JR. Context-Aware Caries Segmentation in Periapical Radiographs Using a Hybrid Multi-Task Learning Framework with Partial Annotations. Applied Sciences. 2026; 16(1):264. https://doi.org/10.3390/app16010264

Chicago/Turabian Style

Nava-Pintor, Jesús Antonio, Héctor A. Guerrero-Osuna, Fabián García-Vázquez, Luis F. Luque-Vega, Teodoro Ibarra-Pérez, Salvador Ibarra-Delgado, Víktor I. Rodríguez-Abdalá, Remberto Sandoval-Arechiga, and José Ricardo Gómez-Rodríguez. 2026. "Context-Aware Caries Segmentation in Periapical Radiographs Using a Hybrid Multi-Task Learning Framework with Partial Annotations" Applied Sciences 16, no. 1: 264. https://doi.org/10.3390/app16010264

APA Style

Nava-Pintor, J. A., Guerrero-Osuna, H. A., García-Vázquez, F., Luque-Vega, L. F., Ibarra-Pérez, T., Ibarra-Delgado, S., Rodríguez-Abdalá, V. I., Sandoval-Arechiga, R., & Gómez-Rodríguez, J. R. (2026). Context-Aware Caries Segmentation in Periapical Radiographs Using a Hybrid Multi-Task Learning Framework with Partial Annotations. Applied Sciences, 16(1), 264. https://doi.org/10.3390/app16010264

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop