Search

The primary objective of this effort is to develop software that will manually and automatically extract traditional anthropometric measurements from 3D point cloud data collected by a whole body scanner for design applications and to define algorithms to statistically summarize the results.

The following data were collected at Fort Jackson, SC during the month of July 1992. The Chief of the CIIP, Mr. Lonnie Turner, guaranteed that all measuring that month would be done by trained CIIP employees, not by soldiers on detail. This was an attempt at getting the most accurate measurements possible. A disclaimer is in order, however, since the standard practice at the CIIP is to round up torso circumference measurements to the next integer. For example, a person measuring 42.25 in the chest was recorded as measuring 43. This imprecision caused the project team to have little confidence in the accuracy of the data. The data were important, however, in that the case-based-reasoning software had to have cases in order to build decision parameters. These were the data used to seed the Remind software tool. Their purpose was to test the usefulness of a case- based-reasoning tool for the task of size prediction in military uniforms (a stable product). The project team hoped to reseed the tool with scan data before the close of the project, but the scanner was not ready. In order for the size prediction tool to be useful, equivalent cases using scan data should be used to seed Remind%. Using the knowledge gained in the learning curve studies (for determining the necessary number of cases needed for useful prediction), a rule-based size prediction scheme, which runs faster than case-based, could then be developed.

The current manual system for selecting uniform sizes for try on is inefficient due to errors in measurement of soldiers and repeated trials of various sizes of garments. Inaccuracies in measurement generally result from the improper placement of the measuring tape and from variations in its tension. This is exacerbated by the use of soldiers on detail assigned the task of measuring. Even among experienced fitters, measurements are inconsistent-measurements taken by the same fitter may vary in consistency when the fitter gets tired. Although size prediction charts are available, they are not effectively used because of the fitters' lack of confidence in the accuracy of the measurements and the difficulties of accounting for priorities among body dimensions. Instead of relying on the prediction charts, a 'you look like this size' approach is generally employed.