Question regarding the calculations of velocity confidence

[paula-tataru]

Hi,

I was looking into how the velocity confidence is calculated and I noticed that the mean velocity is subtracted twice. The first step is to remove the mean per cell across all velocity genes.

V -= V.mean(1)[:, None]

Then, when the confidence for a cell is calculated, the mean per neighbor is again subtracted from the already adjusted velocities.

Vi_neighs = V[indices[i]]
Vi_neighs -= Vi_neighs.mean(1)[:, None]

What is the reason for subtracting the mean twice in the neighbors' velocities?

[WeilerP]

@VolkerBergen, any thoughts on this?