Akademik Veri Yönetim Sistemi
DDS Global Congress 2025 | joint IADMFR Meeting, Venice, İtalya, 16 - 18 Ekim 2025, ss.7-8, (Özet Bildiri)
Abstract:
Sex Classification Based on Radiomics Features of Masseter and Anterior Temporalis Muscles in Ultrasound Images: A Machine Learning Approach
Purpose:
This study aims to develop machine learning models to classify sex based on radiomic features extracted from ultrasound images of the masseter and anterior temporal muscles.
Materials and Methods:
The study included ultrasound images of a total of 174 volunteers (87 females, 87 males), comprising 348 masseter and 348 anterior temporal muscles from both left and right sides. Manual segmentation was performed using 3D Slicer. During image preprocessing, fine and coarse Laplacian of Gaussian (LoG) filters as well as high- and low-pass wavelet transform filters were applied. Radiomic feature extraction was performed using the pyRadiomics extension. Feature selection was conducted using the VarianceThreshold (threshold = 0.8), SelectKBest, and LASSO methods. For sex classification, k-nearest neighbors (KNN), logistic regression (LR), support vector machine (SVM), and random forest (RF) algorithms were employed. Model performance was evaluated using metrics such as area under the receiver operating characteristic curve (AUC), sensitivity, specificity, accuracy, precision, and F1-score.
Results:
The selected features showed statistically significant differences between males and females (p<0.001). The RF and SVM algorithms achieved very high performance in classifying sex.
Conclusion:
Machine learning models developed using radiomic features extracted from ultrasound images successfully classified sex based on the structural characteristics of the masseter and anterior temporal muscles.
Keywords:
Machine learning, radiomics, masseter muscle, anterior temporalis muscle, sex classification, ultrasonography