Camera placement is essential in the world of scientific visualization. Different camera placements expose different information about the data. Viewpoint quality (VQ) metrics are one method of evaluating the quality of camera placement in visualizations. VQ metrics also have the potential to direct the automation of camera selection in scientific visualization, an important issue as the computational capacity to produce data is far outpacing the capacity to save data to storage. Previous research has used VQ metrics to successfully predict camera placements that match user preferences in images generated by a popular visualization technique, called isosurfacing. With this study, we extend the previous research and investigate the efficacy of VQ metrics in predicting user preferred camera placements for images created using another popular visualization method, known as volume rendering.

This study involves two main components: (1) gathering user preferences of camera placements and (2) a performance analysis of how accurately VQ metrics were able to predict user preferences. We found that the top performing VQ metric was able to correctly predict user preferences of volume rendered images up to 66% of the time. This result supports previous findings about the efficacy of VQ metrics in predicting user preferred images generated using isosurfaces. Together, these findings provide further evidence that VQ metrics are a promising approach for guiding the automation of selecting camera placements for scientific visualization methods.