MIMIX-CXR-JPG missing view positions

[mohkoh19]

Hi, I'm currently trying to separate frontal and lateral images in the MIMIC-CXR-JPG dataset. I tried to use ViewPosition and ViewCodeSequence_CodeMeaning columns in the provided metadata CSV file, but I could not reproduce the same number of frontal/lateral images as in the paper (table 3). There are 15,757 missing values for ViewPosition and I couldn't find a way to determine the view position of the corresponding images. Is it possible to share how the frontal/lateral images were separated in the paper? Regards.