NormalizedCrossCorrelationMean.grad#
- NormalizedCrossCorrelationMean.grad()[source]#
Gradient of the score w.r.t. translation (first three) and rotation (last three).
Has the same form as
NormalizedCrossCorrelation.grad()evaluated on the centred valuesvc = v - mean(v)and weightswc = w - mean(w); the mean-subtraction contributes no extra terms becausesum(vc) = sum(wc) = 0(sod<vc, wc>/dp = <dv/dp, wc>andd||vc||/dp = <vc, dv/dp> / ||vc||). Returnsscore_sign * d(score)/dp.