Abstract:
Predictive biomechanical models of the foot need to accurately represent individual-specific surface geometry and underlying anatomy. This is especially an issue for diabetes patients who have foot deformities due to neuropathy and/or amputations. In such cases the biomechanics of the foot is generally severely compromised and accurate analysis of loading is critical for optimising subsequent treatment and development of appropriate orthotic devices. While detailed anatomical information can be obtained from the analysis of magnetic resonance (MR) images of individuals, such approaches are too time-consuming and too costly to be used routinely in the clinic. In order to characterise geometry changes in the diabetic foot in the clinical setting we have constructed a multi-camera stereoscopic device that enables rapid, accurate, and inexpensive measurement of an individual's foot surface geometry. Artificial features were projected onto the foot, using a pseudorandom coloured pattern based on Morano's structured light routine, increasing the yield of stereoscopic data. Neuropathy leads to loss of both sensation and control. This has significant consequences for diabetes sufferers who characteristically develop neuropathy, especially peripheral neuropathy in the feet. The loss of sensation removes the primary mechanism that alerts individuals to foot damage and infection - pain. Nerve death also removes extrinsic control of foot muscles, resulting in further tissue and bone damage due to gait pathologies. Soft tissue inflammation often results from such infections and injuries. Although there is no pain felt by the individual, these regions of tissue inflammation are often characterised by marked regional variations in skin surface temperature. Current measurement systems monitor relatively few sites on the foot, or measure the temperature at the interface between the foot and a rigid platform. We have included an infrared camera in the stereoscope, providing a three-dimensional view of the skin surface temperature distribution over the plantar surface of the foot. The accuracy of the stereoscope was tested by imaging both curved and at surfaces, under a structured light pattern. Least-squares fits produced RMS errors between points and fitted surfaces under 100μm. Geometric data points were fitted to a three-dimensional model of the foot, producing an RMS error of 0.39mm. Two dimensional temperature data were fitted to the geometric model and visualised using a coloured spectrum in a graphical user interface.