I'm testing different methods of calculating missing landmarks for geomorphic morphometric analysis of rodent footprints. After an initial literature review, I came across Bayesian PCA (BPCA) and a least squares regression (LSQR). After calculating these, in R using the PCA wrapper functions in pcaMethods package for BPCA and the best.reg() for LSQR, I am left with a 2D matrix (I believe nxm), but require a 3D array (pxkxn) for all Procrustes analysis. Unfortunately I am struggling to do this step, any advice?

Note: I've tried arrayspecs() in geomorph and vecx() in Morpho. vecX() produced a matrix, but then Procrustes analysis failed to work due to an error with pcAlign.