face recognition - Generalized eigenvalue for two matrices in OpenCV -
face recognition - Generalized eigenvalue for two matrices in OpenCV -
hello working on project involving in face recognition using linear discriminant analysis(lda). lda demands find generalized eigen vectors between class scatter matrix , in class scatter matrix , struck. using opencv devc++ coding. problem looks
a*v=lambda*b*v
where , b matrices generalized eigen vectors should found lambda eigen values , v vectors
upon searching problem many people suggested go calculating inverse of b , multiplying a*v
(inv(b)*a)*v=lambda*v
and calculate eigen vectors inv(b)*a.
it seems solution in case scatter matrix b sigular. found determinant in order of 10^-36 .so cant find inverse , proceed above solution. can 1 suggest me way out of problem except saying code generalized eigen value problem separately.
i providing fisherfaces implementation in github repository @ https://github.com/bytefish/opencv/tree/master/lda. includes implementation of eigenvalue solver general matrices, see: https://github.com/bytefish/opencv/blob/master/lda/include/decomposition.hpp (i've ported great jama solver), exactely looking for.
if have problems code, please drop me note on projects page @ http://www.bytefish.de/blog/fisherfaces_in_opencv.
opencv face-recognition eigenvector eigenvalue
Comments
Post a Comment