Comparison of surface-based and image-based quality metrics for the analysis of dimensional computed tomography data

This paper presents a comparison of surface-based and image-based quality metrics for dimensional X-ray computed tomography (CT) data. The chosen metrics are used to characterize two key aspects in acquiring signals with CT systems: the loss of information (blurring) and the adding of unwanted information (noise). A set of structured experiments was designed to test the response of the metrics to different influencing factors. It is demonstrated that, under certain circumstances, the results of both types of metrics become conflicting, emphasizing the importance of using surface information for evaluating the quality dimensional CT data. Specific findings using both types of metrics are also discussed.