Skip to main content
. Author manuscript; available in PMC: 2021 Mar 30.
Published in final edited form as: IEEE Trans Med Imaging. 2021 Mar 19;PP:10.1109/TMI.2021.3067512. doi: 10.1109/TMI.2021.3067512

Figure 1.

Figure 1

The preferences of the observers for each of the reconstruction methods: BASELINE, ZSSR and SISR. The preference is shown as a percentage of the choices calculated as a number of choices for each method from total possible choices in the survey for an expert (1a) and a non-expert (1b).

(a) Preferences of expert observers. Majority of experts has a clear tendency towards ZSSR over BASELINE, with the lowest number of choices for SISR. Experts 2 and 4 has a dominant preference towards SISR over ZSSR reconstructions, which is skewed to the rest of observers. Thus, observers 2 and 4 need more attention and are treated as outliers in further analysis.

(b) Preferences of non-expert observers. There is a clear preference for SISR reconstruction for the majority of observers. In contrast, no dominant preference occurs for neither ZSSR and BASELINE. In comparison to others, observers, e.g. 4, 10, 15, 19, 27, 28, 31, 35 do not show a preference for SISR; they should be investigated as statistically valid outliers.