Three-dimensional deep-learning age and sex estimation from the fourth rib: de-identified CT-derived surface dataset and reference implementation

453 de-identified right fourth-rib surface meshes (binary STL, millimetres) segmented from clinical thin-section thoracic CT of adults aged 18-100 years (225 men, 228 women), labelled with sex and age only, with a frozen development/test split, 30 repeat segmentations for intra-observer analysis, and the complete deep-learning pipeline (multi-view RGB-D CNN, PointNet++, morphometric scalars, label-distribution learning, calibrated likelihood-ratio reporting) used in the accompanying Forensic Science International article. Data: CC BY 4.0; code: MIT.