forensic-anthropology

SexEst is an open-source Streamlit web application for predicting biological sex from skeletal measurements using machine learning (XGBoost, LightGBM, Linear Discriminant Analysis). The best-performing models achieved cross-validated accuracies of ~80–90% on the Goldman (postcranial) and Howells (cranial) datasets.

Osteometric sorting for commingled human remains

Jupyter notebooks for training ML models that predict biological sex from skeletal measurements. Companion to the SexEst web app.

ANP 892 Advanced Osteology Project

Files and code for the AABA 2022 Presentation: "Advances in Subadult Age Estimation: Evaluating the Performance of the Mixed Cumulative Probit (MCP) on Contemporary Subadults"

Interactive Dash app for skeletal age‑at‑death estimation using pre-trained ML models (classification + regression).

Interactive 3D forensic face approximation from skull measurements, using real tissue depth data and RBF interpolation. Built with Python, SciPy, and Plotly Dash.

Society of Forensic Anthropologists Forensic Case Database

AgeEst notebook: reproducible analysis used to build the AgeEst models (DOI: 10.1016/j.fsir.2023.100317). Full dataset not included; contact authors for access.

"Measured quality reports & verification tools for CT-derived skull mesh assets"

Often comparisons of the theoretical and anatomical basis, data collection methods, reliability and success rates, as well as details of the sample population used to establish these guides cannot be readily made due to their heterogeneity, leaving them to interpret these recommendations and assess their limitations.

Population-specific sex estimation from postcranial metrics (ancient Dion, Greece): Colab notebook + exported Logistic Regression formulas and ML models (XGBoost/LightGBM/RF), with imputation and cross-validation comparisons