Sex estimation - long bones

Sex estimation from twelve osteometric measurements taken on the humerus, radius, femur and tibia, using a machine-learning model (linear discriminant analysis) validated on three reference samples.

Humerus

	Value
humxln	mm
humhhd	mm
hummwd	mm

Radius

	Value
radxln	mm
radtvd	mm
radapd	mm

Femur

	Value
femxln	mm
femebr	mm
femhhd	mm
femmtv	mm

Tibia

	Value
tibxln	mm
tibpeb	mm

Decision threshold

0.50

Below this probability, the tool does not decide. Raising the threshold raises the correct classification rate and lowers the proportion of individuals assigned.

The measurements entered are sent to a third-party calculation server, from server to server: your browser contacts no third party. Nothing is stored or logged, on either side.

Measurement definitions

Upper limb

Diagrams of the measurement points on the humerus (maximum length, epicondylar breadth, head diameter, midshaft diameter) and the radius (maximum length, transverse and antero-posterior diameters at midshaft).

Bone	Code	Measurement
Humerus	humxln	Maximum length
	humhhd	Maximum vertical diameter of the head
	hummwd	Minimum midshaft diameter
Radius	radxln	Maximum length
	radtvd	Transverse diameter at midshaft
	radapd	Antero-posterior diameter at midshaft

Lower limb

Diagrams of the measurement points on the femur (maximum and physiological length, epicondylar breadth, head diameter, midshaft medio-lateral diameter) and the tibia (length, proximal epiphyseal breadth).

Bone	Code	Measurement
Femur	femxln	Maximum length
	femebr	Epicondylar breadth
	femhhd	Maximum diameter of the head
	femmtv	Medio-lateral diameter at midshaft
Tibia	tibxln	Lateral condyle to medial malleolus length
Tibia	tibpeb	Maximum breadth of the proximal epiphysis

Training collections

The model was trained on a combined database of ten skeletal collections. The performance figures shown after an estimate come from three groups evaluated separately: Milan and Pretoria (marked below), each evaluated individually, and "BDC" (marked below), which combines the eight other collections evaluated together.

Group		Country	Period	Sex	Females	Males	Total
BDC	Goldman	International	Antiquity - 21st century	Estimated	543	985	1528
	Laudun	France	5th - 13th century	Estimated	54	113	167
	La Ciotat (LC)	France	16th - 19th century	Estimated	105	109	214
	Marseille Petite Crotte (MPC)	France	18th - 20th century	Estimated	57	47	104
	Terry	United States	20th century	Known	207	247	454
	Olivier	France	20th century	Known	40	62	102
	Nice	France	21st century	Known	21	19	40
	NMDID (imagerie CT)	United States (New Mexico)	2010 - 2017	Known	50	50	100
validation	CAL Milan	Italy	21st century	Known	199	199	398
validation	Pretoria Bone Collection (PBC)	South Africa	21st century	Known	180	180	360
Total					1456	2011	3467

References

Knecht, S., Santos, F., Ardagna, Y., Alunni, V., Adalian, P., & Nogueira, L. (2023). Sex estimation from long bones: a machine learning approach. International Journal of Legal Medicine, 137(6), 1887-1895. https://doi.org/10.1007/s00414-023-03072-4

Knecht, S., Krüger, G., Liebenberg, L., Ardagna, Y., Perrin, M., Ouladsine, M., ... & Adalian, P. (2026). Case-specific accuracy in sex estimation from long bones in forensic anthropology: an "Accuracy x-Factors" approach. Forensic Science International, 112820. https://doi.org/10.1016/j.forsciint.2026.112820

Knecht, S., Morandini, P., Biehler-Gomez, L., Ardagna, Y., Perrin, M., Cattaneo, C., ... & Adalian, P. (2026). Interpretable machine learning for individualized sex estimation from long bones. International Journal of Legal Medicine, 140(2), 983-995. https://doi.org/10.1007/s00414-025-03635-7

Knecht, S., Morandini, P., Biehler-Gomez, L., Ouladsine, M., Roman, C., Cattaneo, C., & Adalian, P. (2026). Population-specific validation of long bone linear discriminant analysis against morphoscopic pelvic methods for sex estimation in contemporary Italian skeletal remains. Journal of Forensic Sciences, 71(3), 1388-1397. https://doi.org/10.1111/1556-4029.70293