How to understand the J_helper represention as follows?
I tried to understand the following code in so3_spline.h(line 233 to 268), however, after some calcilations(according the paper Efficient Derivative Computation for Cumulative B-Splines on Lie Groups), I found that my anser is so different as yours, did I make something wrong?
J_helper = coeff[i + 1] * res.matrix() * Jl_k_delta * Jl_inv_delta * p0.inverse().matrix();
J->d_val_d_knot[i] -= J_helper;
What I have in mind is that: coeff[i + 1] coresponds \lambda_j and kdelta equals \lambda_j \mathbf{d}_j as same in the paper quoted above, so am I missimg something?