Alignment of plantar pressure data with 3D foot placement data
H Davies, P Davenport
Clinical Measurement Laboratory, Birmingham Community Healthcare NHS Trust
Introduction
Measurement of pressure on the plantar surface of the foot is recognised as being a useful tool in characterising the effect of orthotic intervention and in evaluating at-risk sections of the foot. Plantar pressure measurement devices can be broadly divided into walkways (where the subject walks on a surface containing sensors, covering one or more footsteps) or insole devices (worn within footwear – better able to assess orthotic changes). An example of the latter, the FScan (Tekscan, USA) is currently used by the Clinical Measurement Laboratory.
Problem Definition
Peak pressure and pressure integral values and the trajectory and
variation of the centre of pressure during each stance can be of interest to referring clinicians. Foot progression angle is known to affect plantar pressure measurements (Chang et al (2006)), and provides context to the reported values. Insole systems are not able to independently report foot progression angle.
Simultaneous measurement of plantar pressure and foot progression
angle has been reported previously. Giacomozzi et al (1998) presented details of a combined pressure mat-force plate-2D video tracking system. Sawacha et al (2012) reported on a kinematics-kinetics-plantar pressure data collection. The literature review was unable to find reference to synchronised measurement of foot progression angle and in-shoe plantar pressure maps.
Project Aim
The aim of this project was to develop a means of incorporating the foot
progression angle (as measured using a 3D motion capture system) with the images of peak pressure and centre of pressure trajectory. By synchronising these, the effect of foot progression angle on the plantar pressure
recordings can be evaluated. In effect the integration of these systems allows the insole system to gain some of the benefits of a
walkway device. The information generated would be particularly useful to orthotists and podiatrists that refer patients to the Clinical Measurements Laboratory for assessment.
Data Collection and Processing
Retro-reflective markers were placed on the heel and on the dorsal surface of the shoe at approximately the position of the
2-3rd metatarsal heads bilaterally on a single unimpaired subject. Position data were collected using a 12 camera VICON Nexus (Oxford Metrics, UK) motion capture system. FScan in-shoe pressure measurement sensors were cut to size and taped securely inside the subject’s footwear. The VICON system was calibrated according to the standard laboratory procedures, the FScan system using a ‘Step’ calibration to the subject’s body weight as recommended by the manufacturer.
To merge the data from the FScan system with the Vicon position data, custom software, F-RAT (FScan Rotational Alignment Tool) was written using Matlab (Mathworks, USA). Images of the peak stance pressures were taken for each stance from the FScan software using Snagit (Techsmith, USA) (figure 1). The toe and heel markers for each foot were labelled using Vicon Nexus (figure 2). The heel strike and toe off events were also identified and saved in a c3d file.
The c3d server (Motion Lab Systems, USA) was used in Matlab to read the marker positions and events from the c3d file. The angle of the toe-heel vector was calculated at each midstance, and the plantar pressure image rotated to match this angle (figure 3). A series of steps were then plotted, such that the heel position on the image was at the heel position in the laboratory coordinate system. The resulting figure (figure 4) could then be printed for inclusion in a report.
Discussion
One means of validating the alignment of the long axis of the foot with the vector of the heel-toe markers is to use a force platform. By performing a principal component analysis of both the centre of pressure trajectory on the insole and of the centre of force path recorded by the forceplate, a correction to the foot progression angle can be included. Currently this method assumes that they are collinear.
The software assumes that the foot progression angle is constant throughout midstance. Normative data previously collected suggests this is not an unreasonable assumption, but it may be clinically useful to examine changes at early and late stance. By rotating pressure video at each frame, this detail can be reported. This presents greater difficulty in synchronisation and processing.
Before being used as a clinical tool, the effect of marker misplacement and the repeatability of clinicians using the technique would need to be assessed. Medial/lateral displacement of the toe and heel marker alters the reported angle: consistency in locating and marking the appropriate position is essential. This requires assessment, training and regular repeatability audit.
Conclusions
The system presented here is a novel method of combining in-shoe pressure measurement with the spatial information typically available from a pressure mat. This allows an easy visual comparison of left and right, assessment of foot placement and better understanding of the recorded pressure distribution. Further work is underway to automate the extraction of stance images and synchronise capture between devices for applying F-RAT directly to pressure videos.
References
W Chang et al, “Impact of changing foot progression angle on foot pressure measurement in children with neuromuscular diseases”, Gait and Posture, vol. 20, p14-19, 2004.
C Giacomozzi et al, “Integrated force-pressure-position measurements for the in-vivo characterisation of the plantar foot loading” 11th Conference of the European society of biomechanics, p132, 8th July, 1998.
Z Sawacha et al, “Integrated kinematics-kinetics-plantar pressure data analysis: A useful tool for characterising diabetic foot biomechanics”, Gait and Posture, vol. 36, p20-26, 2012.