On a new method to assess the accuracy of a CAD program

This study was initiated with the purpose of developing and evaluating a system for measuring the magnitude of the variation between a computer-aided design (CAD) object created on the computer screen and a replicated object produced by computer-aided manufacturing (CAM).

Controlled geometric forms, a square and a cone, were designed in a CAD program, and measuring distances were selected. The CAD data were transmitted to CAM data, and objects were produced. The CAD/CAM process was the Procera system, and the holder system permitted the objects to be ground into cross sections, similar to the visualization in the CAD program. Five objects of each shape were produced and measured by two operators two times. Three operators measured one of the objects 30 times.

Observed values were greater than the true value. For all objects, the systematic error was at most 15.5 microns. Interoperator difference was small. The variation because of measurement error was greater for the square object compared to the cone. However, the variation because of object was higher for the cone object than for the square. The total standard deviation was 7.7 microns. Thus, the total random error caused by object variation and measurement error was in approximately 95% of all measurements less than 15 microns.

There are no differences in the measurement data derived from this method and actual measurement data from an object created by the computer-aided dental design program. The method has high validity and reliability, i.e., high accuracy.