Variability in assessing visual image quality

In diagnostic radiology, the variability in image quality evaluation has been widely recognized as a common phenomenon. Accordingly, a noticeable body of research has been created to address this phenomenon and improve consistency in relation to dose levels and image quality. In the context of image quality assessment, system performance may not only be the sole reason behind diagnostic variations. Observer variability could also have a significant contribution on the overall diagnostic accuracy. This issue could influence the reliability of the results obtained from visually based image quality assessment methods. Variability in determination of image quality has been investigated since the 1940s. This variation in image quality assessment may result from a lack of standards, including those in reference to visual grading scales. In this context, Krupinski and Jiang (2008) have identified two important issues which need to be considered when addressing variability:

1) that systems are required to minimize the variation between observer interpretations;

2) that approaches are required to assess the systems and their influence on observer interpretation.

The European Guidelines for image quality criteria (CEC, 1996) could be one measure used to address the variability in image quality assessment.