search

ClaireLinHong / regularized-field-map-estimation

5 stars 1 forks source link

RMSE in Hz or not for Fig. 4? #5

Closed JeffFessler closed 1 year ago

JeffFessler commented 1 year ago

Fig. 4 in the paper shows RMSE labeled in Hz: https://github.com/ClaireYLin/regularized-field-map-estimation/blob/1e60f83feb7baafd5d767e777f8fd68033488ea3/examples/example_simu_fieldmap.m#L126 but the data passed to compute_rmsd is all "w" arrays like wtrue in units of rad/s https://github.com/ClaireYLin/regularized-field-map-estimation/blob/1e60f83feb7baafd5d767e777f8fd68033488ea3/examples/example_simu_fieldmap.m#L122 So perhaps the actual RMSE values are 2\pi lower than what is shown in Fig. 4, right? Please just reply without changing code because I am in the middle of making several changes.

JeffFessler commented 1 year ago

A closely related question is what is the factor of 2 in this RMSD calculation: https://github.com/ClaireYLin/regularized-field-map-estimation/blob/b8f0e7faec1fb5c6974bc9cd9caf946a6190d582/functions/compute_rmsd.m#L27 Was that supposed to be 2*pi? If so our graphs are probably off by pi not by 2pi.

ClaireLinHong commented 1 year ago

You are right - the input to compute_rmsd is wtrue, but the error should be divided by 2*pi instead of 2.