Comparison of body segment models for female high jumpers utilising DXA images

In motion analysis of sport competitions, the question is often about the most convenient choice for defining the segment endpoints when no visible landmarks can be used. The purpose of the present study was to determine the location of the body centre of mass (CoM) of female high jumpers by using a high accuracy reaction board and two different segment models: Dempster (1955) and de Leva (1996). Digitising the bony landmarks from the images of dual energy x-ray absorptiometry (DXA) and overhead digital camera were used to compare the digitising accuracy. The location of the CoM determined by a reaction board was 55.88 ± 0.52% of subjects' body height. The segment model of Dempster digitized from DXA images (56.66 ± 0.50%) differed from the reference values of reaction board (p = 0.004), whereas the model of de Leva (56.06 ± 0.61%) showed no significant difference. The model of de Leva adjusted for female subjects differed only slightly (0.32%), thus, providing appropriate model for female high jumpers. Since the digitised bony landmarks in the DXA images are obviously very close to the correct locations, the differences in results between the segment models and reaction board is most likely due to inaccuracies in the model itself and/or generalisation of one model to different body structures. When the segment landmarks were estimated without any markers on the body, the results did not differ much from the DXA results.