HERMES was developed to fill a need for a fast FE model to conduct the thousands of simulations needed to optimize protection systems for populations. Most FE models represent the same two or three sizes of people (the so-called 5th, 50th, and 95th percentile). But humans have a wide range of shapes and sizes, so accurately representing a population requires a parametric model that can rapidly generate geometry for any person.
HERMES is under continuous development, but we are sharing it publicly to encourage researchers to consider population diversity when designing protection systems. The non-commercial license allows for use in internal research and development only. Use in academic research is permitted. If you would like to use HERMES in a commercial application (that is, as part of a product or service for sale), please contact us.
Most importantly, please let us know how you are using the model and contact us with any problems or suggestions for improvements.
If you publish research using HERMES, please cite:
Hu J, Dhawan M, Fan X, and Reed MP (2022) Toward Improved Safety Equity: An Efficient, Parametric, and Posable Finite Element Human Model Representing a Diverse Population, IRCOBI Europe Conference. http://www.ircobi.org/wordpress/downloads/irc22/pdf-files/2237.pdf
The following Figure shows the general methods for developing the parametric HERMES models. Statistical geometry models of skeleton (skull, cervical spine, ribcage, lumbar spine, pelvis, femur and tibia) and external body surface are used to predict the geometry for the target body dimensions based on age, sex, stature, body mass index (BMI), and sitting height to stature ratio (SHS). Such statistical geometry models were developed in our laboratory over the past 10 years based on clinical CT scans and volunteer body scans from hundreds of subjects through image segmentation, landmarking, mesh morphing/mapping, Generalized Procrustes Alignment (GPA), Principal Component Analysis (PCA), and regression analysis. Bony landmarks collected from the body scan data were used to integrate the skeleton into the body surface model, while local mesh morphing was performed to correct potential bone-to-skin penetrations. A mesh morphing method links the baseline HERMES mesh to the statistical geometry targets, so that the morphed models can represent individuals with a wide range of body size and shape.
The following Table shows the subject sample size and associated publications for developing the statistical models of different bones and body shape. A brief introduction of HERMES development and validations can be found in this IRCOBI paper.
|Body Region||Sex||Subject Number
(CT & Body Scans)
|External Surface||F||73||(Park et al. 2021)|
|Pelvis||F||75||(Brynskog et al. 2021)|
|Femur||F||36||(Klein et al. 2015)|
|Ribcage||F||47||(Wang et al. 2016) Recently updated|
|Cervical Spine||F||79||(Reed and Jones 2017)*|
|Lumbar Spine||F||82||(Tang et al. 2022)*|
|Skull||F||54||(Wei et al. 2022)|