Fig. 3
From: Validity of two subjective skin tone scales and its implications on healthcare model fairness
Data Pipeline and Protocol Flowchart. The data collected using the GUI was stored as separate spreadsheets for each annotator and each scale. Organizing this data starting with consolidating all of these spreadsheets into a Pandas Dataframe to start systemic evaluation across annotator ratings and scales. The patient subject data was also stored in a separate spreadsheet which was linked to each image using the common patient id number. All data organization and statistical analysis was conducted in Python. The Fitzpatrick (scale of I–VI) and Monk (scale of 1–10) refer to the two skin tone scales used for this study.
