The measurement of limb volume and shape is an important tool in vascular surgery. The limb volume can change by as much as 30% when a patient changes from a prone to a vertical position. The current measuring technique of immersing the limb in water and measuring the displacement has an accuracy of ± 15%. This method is not sufficiently accurate and is difficult to administer with little scope to measure the changes in volume with orientation for modern medical/surgical procedures.
The Image Processing Laboratory at the University of Limerick has for many years applied machine vision techniques to the solution of inspection problems in difficult to measure castings and machined components. In co-operation with the Vascular Surgery Unit at the Regional Hospital Limerick these same machine vision methods were used to measure limb volume. Machine vision in this application is non-invasive, non-contacting and applicable to any limb orientation. There is also an opportunity to generate three-dimensional graphical representations of the limb with the system.
In the adopted method a plane of light is projected onto the limb from three separate sources located at 120° intervals around the limb. Three cameras image the intersections of these light sources with the limb. The views from the cameras are merged to generate a complete 360° contour, which is effectively a cross-section of the limb at that point. The method also successfully records shallow concave regions.
To obtain the volume, the limb is scanned to give a series of closely spaced cross-sections which are summed to give the volume over the length measured. In preliminary studies accuracies of greater than ± 5% have been achieved. The research is progressing and it is expected that with the use of higher order polynomials accuracies in the order of ± 0.5% will be achieved.