OBJECTIVE: The shape of the current physical and computational surrogates of children used for restraint system assessments is based largely on standard anthropometric dimensions. These scalar dimensions provide valuable information on the overall size of the individual but do not provide good guidance on shape or posture. This study introduced the development of a parametric model that statistically predicts individual child body shapes in seated postures with a few given parameters. METHODS: Surface geometry data from a laser scanner of children ages 3 to 11 (n = 135) were standardized by a 2-level fitting method using intermediate templates. The standardized data were analyzed by principal component analysis (PCA) to efficiently describe the body shape variance. Parameters such as stature, body mass index, erect sitting height, and 2 posture variables related to torso recline and lumbar spine flexion were associated with the PCA model using regression. RESULTS: When the original scan data were compared with the predictions of the model using the given subject dimensions, the average root mean square error for the torso was 9.5 mm, and the 95th percentile error was 17.35 mm. CONCLUSIONS: For the first time, a statistical model of child body shapes in seated postures is available. This parametric model allows the generation of an infinite number of virtual children spanning a wide range of body sizes and postures. The results have broad applicability in product design and safety analysis. Future work is needed to improve the representation of hands and feet and to extend the age range of the model. The model presented in this article is publicly available online through HumanShape.org.
OBJECTIVE: The shape of the current physical and computational surrogates of children used for restraint system assessments is based largely on standard anthropometric dimensions. These scalar dimensions provide valuable information on the overall size of the individual but do not provide good guidance on shape or posture. This study introduced the development of a parametric model that statistically predicts individual child body shapes in seated postures with a few given parameters. METHODS: Surface geometry data from a laser scanner of children ages 3 to 11 (n = 135) were standardized by a 2-level fitting method using intermediate templates. The standardized data were analyzed by principal component analysis (PCA) to efficiently describe the body shape variance. Parameters such as stature, body mass index, erect sitting height, and 2 posture variables related to torso recline and lumbar spine flexion were associated with the PCA model using regression. RESULTS: When the original scan data were compared with the predictions of the model using the given subject dimensions, the average root mean square error for the torso was 9.5 mm, and the 95th percentile error was 17.35 mm. CONCLUSIONS: For the first time, a statistical model of child body shapes in seated postures is available. This parametric model allows the generation of an infinite number of virtual children spanning a wide range of body sizes and postures. The results have broad applicability in product design and safety analysis. Future work is needed to improve the representation of hands and feet and to extend the age range of the model. The model presented in this article is publicly available online through HumanShape.org.
Entities:
Keywords:
anthropometry; body shape prediction; child body shape measurement; parametric modeling; seated body shape; statistical body shape model
Authors: Laura Medialdea; Barry Bogin; Mbeugue Thiam; Antonio Vargas; María D Marrodán; Nicole I Dossou Journal: Sci Rep Date: 2021-02-19 Impact factor: 4.379