Question about experiment result

[Lancial]

Hi! Could you clarify how was the initial average landmarks error of the DIRLab dataset without registration calculated? Specifically, in your paper, you wrote
"The initial average landmark distance between consecutive phase images was 2.25 ± 1.55 mm for the Popi datasets and 2.19 ± 2.02 mm for the DirLab datasets"

I think this number was also shown in Fig.4 left.

I could not reproduce this number with the DIRLab dataset.

[lxy51]

Hello, I would like to ask, how do you get the data after the author preprocessed? Could you please share with me a copy of the pre-processed dirlab database