Search

A new system for measuring the cross-sectional area profiles of amputation stumps and whole limbs has been designed at the Amputee Research Centre. The instrument consists of a cylindrical tank supported on an elevator. The tank is raised to the height of the amputation stump and filled with water. A graph of the cross-sectional area profile of the amputation stump is generated by a mini-computer as the elevator descends. The cross-sectional area (A) is calculated from the expression: formula: (see text) where Hw = height of water in the tank He = height of the elevator Ac = a constant, related to the size of the measuring tank. This paper describes the instrument, which may find application in many other areas where there is a need to study shape.

The single polynomial analysis of ionograms is adapted to give a least-squares procedure in which the number of scaled virtual heights is greater than the number of terms in the real height expansion. More terms can then be used in the polynomial without incurring unphysical oscillations. The detail required in the real height profile is set by the number of terms independently of the number of scaled virtual heights. Any mixture of ordinary and extraordinary ray virtual heights can be used. The accuracy with which the calculated profile fits the observations is given by an rms fitting error; the accuracy can also be checked at each scaled frequency, to detect (and reject) incorrect points. The method is particularly appropriate for the analysis of topside ionograms since full use can be made of fragmentary ordinary and extraordinary ray traces. These are incorporated into a single analysis which produces a real height curve interpolating smoothly across any unobserved regions. An 8 or 10 term polynomial, fitted to 10 to 20 virtual heights, is suitable for routine work. The resulting analytic expression can give any required number of points on the real height profile.

A percutaneous implant for the attachment of an artificial limb has been designed and tested in 14 pigs. Firm fixation to beone was achieved with the porous-surface layered metal intramedullary stem design in some cases. Dacron velour was used at the soft tissue interface. Evidence of soft tissue ingrowth was seen. However, the velour was unable to maintain adequate epithelial adhesion to form an anatomical seal and a barrier to bacteria.

The phenomenon of aliasing in computer graphics is reviewed, and the antialiasing approach of prefiltering before sampling adopted. The theoretical performance of five different types of convolutional filter is studied, and a rotated Hamming filter is selected as the best. The Edge Spread Function, which forms the basis for the antialiasing scheme, is given for the different filters. A decomposition approach reduces the problem of filtering an arbitrary polygon to the problem of filtering acute-angled segments of coordinate space, called vertex areas. A scheme for filtering vertex areas is developed and shown to have a maximum error, in a small region around the vertex itself, of about 4% of the intensity of the polygon. The scheme is incorporated into a scan-line algorithm for scan-conversion of two-dimensional polygonal scenes; the algorithm is believed to be simpler and faster than any existing antialiasing schemes that use weighted filters.