PURPOSE: The purposes of this study were (i) to develop a three-dimensional interactive visualization tool for exploring plantar pressure time series and spatiotemporal statistical volumes and (ii) to demonstrate the benefits of volumetric analyses using various running and walking data sets. METHODS: A data exploration tool was developed in Python using the open-source Visualization Toolkit. Multiple-pressure isosurfaces were computed and were then rendered with interactive rotation and adjustable thresholds and transparencies. Plantar pressure data were collected: (i) from two running subjects, one with a heel-loading pattern and one with a forefoot-loading pattern; (ii) from one individual while running straight and then while performing a cutting maneuver; and (iii) from one subject walking at three different speeds. All data were spatiotemporally aligned, and mean volumes were computed. Statistical volumes were also computed for the walking data set, and significance was assessed topologically using techniques from three-dimensional brain imaging. RESULTS: After converting raw plantar pressure data into a rapidly readable format, volumetric renderings were presented in ∼50 ms, a negligible time lag for interactive data exploration. We observed that consideration of only spatial two-dimensional variables yielded "impulse illusions" that could be resolved most effectively with three-dimensional renderings. For all data sets, we found that dynamic foot behavior was clearest through interactive three-dimensional exploration. CONCLUSIONS: Plantar pressure data contain high-quality biomechanical information in their original three-dimensional form. The main benefit of the proposed visualization technique is that it affords qualitatively rich and unique holistic explorations of dynamic foot behavior.
PURPOSE: The purposes of this study were (i) to develop a three-dimensional interactive visualization tool for exploring plantar pressure time series and spatiotemporal statistical volumes and (ii) to demonstrate the benefits of volumetric analyses using various running and walking data sets. METHODS: A data exploration tool was developed in Python using the open-source Visualization Toolkit. Multiple-pressure isosurfaces were computed and were then rendered with interactive rotation and adjustable thresholds and transparencies. Plantar pressure data were collected: (i) from two running subjects, one with a heel-loading pattern and one with a forefoot-loading pattern; (ii) from one individual while running straight and then while performing a cutting maneuver; and (iii) from one subject walking at three different speeds. All data were spatiotemporally aligned, and mean volumes were computed. Statistical volumes were also computed for the walking data set, and significance was assessed topologically using techniques from three-dimensional brain imaging. RESULTS: After converting raw plantar pressure data into a rapidly readable format, volumetric renderings were presented in ∼50 ms, a negligible time lag for interactive data exploration. We observed that consideration of only spatial two-dimensional variables yielded "impulse illusions" that could be resolved most effectively with three-dimensional renderings. For all data sets, we found that dynamic foot behavior was clearest through interactive three-dimensional exploration. CONCLUSIONS: Plantar pressure data contain high-quality biomechanical information in their original three-dimensional form. The main benefit of the proposed visualization technique is that it affords qualitatively rich and unique holistic explorations of dynamic foot behavior.