@article{ART003286275},
author={Kyeongsu Lee and IZZATI LIA WILDA and Yeji Kim and Mun ingyeong and Jo Byungdu and Yongsu Yoon},
title={Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology},
journal={Anatomy & Biological Anthropology},
issn={2671-5651},
year={2025},
volume={38},
number={4},
pages={311-318},
doi={10.11637/aba.2025.38.4.311}
TY - JOUR
AU - Kyeongsu Lee
AU - IZZATI LIA WILDA
AU - Yeji Kim
AU - Mun ingyeong
AU - Jo Byungdu
AU - Yongsu Yoon
TI - Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology
JO - Anatomy & Biological Anthropology
PY - 2025
VL - 38
IS - 4
PB - 대한체질인류학회
SP - 311
EP - 318
SN - 2671-5651
AB - Sex estimation is a key procedure in the forensic identification process, and skeletal-based assessment is primarily performed using sexual dimorphism observed in the pelvis and skull. The paranasal sinuses, including the maxillary sinus, are located within the facial bones and tend to remain relatively well-preserved after death. Due to their substantial anatomical variability among individuals, their potential use as an auxiliary indicator in forensic identification has been suggested. This study aimed to determine whether differences in sex and laterality exist by automatically segmenting the maxillary sinus and calculating its volume using a U-Net-based deep learning model applied to 70 Korean postmortem computed tomography scans. As a result, the U-Net model demonstrated overall high segmentation performance (0.8921±0.019). The maxillary sinus volumes calculated by the model showed no statistically significant differences between sexes or between the left and right sides, which is consistent with findings reported in previous studies. This study demonstrates the feasibility of deep learning-based morphological analysis of the maxillary sinus. In the future, by generating a large training dataset using both ante-mortem and postmortem computed tomography images with the proposed model, it is expected that a three-dimensional CNN-based system for sex and age estimation may be developed, potentially achieving superior performance compared with existing approaches.
KW - Post-mortem CT;Maxillary sinus;Deep learning;Morphological analysis;Personal identification
DO - 10.11637/aba.2025.38.4.311
ER -
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu and Yongsu Yoon. (2025). Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology. Anatomy & Biological Anthropology, 38(4), 311-318.
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu and Yongsu Yoon. 2025, "Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology", Anatomy & Biological Anthropology, vol.38, no.4 pp.311-318. Available from: doi:10.11637/aba.2025.38.4.311
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu, Yongsu Yoon "Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology" Anatomy & Biological Anthropology 38.4 pp.311-318 (2025) : 311.
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu, Yongsu Yoon. Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology. 2025; 38(4), 311-318. Available from: doi:10.11637/aba.2025.38.4.311
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu and Yongsu Yoon. "Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology" Anatomy & Biological Anthropology 38, no.4 (2025) : 311-318.doi: 10.11637/aba.2025.38.4.311
Kyeongsu Lee; IZZATI LIA WILDA; Yeji Kim; Mun ingyeong; Jo Byungdu; Yongsu Yoon. Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology. Anatomy & Biological Anthropology, 38(4), 311-318. doi: 10.11637/aba.2025.38.4.311
Kyeongsu Lee; IZZATI LIA WILDA; Yeji Kim; Mun ingyeong; Jo Byungdu; Yongsu Yoon. Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology. Anatomy & Biological Anthropology. 2025; 38(4) 311-318. doi: 10.11637/aba.2025.38.4.311
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu, Yongsu Yoon. Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology. 2025; 38(4), 311-318. Available from: doi:10.11637/aba.2025.38.4.311
Kyeongsu Lee, IZZATI LIA WILDA, Yeji Kim, Mun ingyeong, Jo Byungdu and Yongsu Yoon. "Development of Deep Learning-based Model for Automatic Calculation of Maxillary Sinus Volume Using Post-mortem Computed Tomograpy in Forensic Anthropology" Anatomy & Biological Anthropology 38, no.4 (2025) : 311-318.doi: 10.11637/aba.2025.38.4.311
1