Sign in Subscribe
Machine learning

Machine Learning Algorithms Accurately Determine Sex from CT Images

Strategic Insights

— 1 min read

Researchers from Karabuk University have developed a machine learning model that can accurately determine sex from computed tomography (CT) images of the skull. The study analyzed CT images of 200 individuals, aged 19-65 years, and used various machine learning algorithms to estimate sex based on morphometric measurements from the facial canal. The results showed that all algorithms, except for Quadratic Discriminant Analysis (QDA), achieved an accuracy rate of 0.97, with the multilayer perceptron classifier (MLPC) from ANN algorithms performing best.

The study's findings have significant implications for forensic investigations, particularly in cases where craniofacial remains are fragmented, and for preoperative planning in otologic and skull base surgeries. The research suggests that FN-centered morphometric measurements can aid in understanding facial nerve positioning across sexes and populations, providing valuable diagnostic data for surgeons, anthropologists, and forensic experts.

Key Takeaways:

  • The study utilized CT images of 200 individuals, aged 19-65 years, with 100 females and 100 males.
  • Nine temporal bone parameters were measured in axial, coronal, and sagittal planes.
  • Machine learning algorithms, including Quadratic Discriminant Analysis (QDA), Linear Discriminant Analysis (LDA), Decision Tree (DT), Extra Tree Classifier (ETC), Random Forest (RF), Logistic Regression (LR), Gaussian Naive Bayes (GaussianNB), and k-Nearest Neighbors (k-NN), were used to estimate sex.
  • The multilayer perceptron classifier (MLPC) from ANN algorithms achieved the highest accuracy rate of 0.97.
  • SHapley Additive exPlanations (SHAP) analysis revealed the five most impactful parameters: right SGAs, left SGAs, right TSWs, left TSWs, and the inner mouth width of the left FN.
  • The study's findings have implications for forensic investigations, preoperative planning, and diagnostic data for surgeons, anthropologists, and forensic experts.

Statistics:

  • 200 individuals were included in the study, with 100 females and 100 males.
  • The age range of the participants was 19-65 years.
  • The accuracy rate of the machine learning algorithms, except for QDA, was 0.97.
  • The multilayer perceptron classifier (MLPC) from ANN algorithms achieved the highest accuracy rate of 0.97.
  • The five most impactful parameters were right SGAs, left SGAs, right TSWs, left TSWs, and the inner mouth width of the left FN.

Sources:

  • Sex estimation with parameters of the facial canal by computed tomography using machine learning algorithms and artificial neural networks. BMC Medical Imaging, 2025,25(1):1-13.
  • BMC Medical Imaging - http://bmcmedimaging.biomedcentral.com
  • DOI: 10.1186/s12880-025-01834-7

Read more