Ask Your Question

How to find Eigenvectors of high dimension CovarianceMatrix ?

asked 2017-03-11 10:19:47 -0600

Dashu gravatar image

I am implementing PCA function, so I can not just call PCA function of OpenCV. I have trouble in calculating eigenvectors using Eigen function, input covariance matrix is 10000x10000 (because image size is 100x100, so 10000 features for one image, training set is only 5 images), eigenvectors returned by Eigen is not usual, every row of eigenvectors has 9999 "0" and only one "1", 00000000.....0010000....00000, I think this vector is normalised already. I try to use low dimension input covariance matrix, eigenvectors is ok. I also try to use Eigen function of Eigen library(, but it take a lot of time and I can not wait. Any function from OpenCV I can use? Or I need to implement my own Eigen function to reduce cost?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2018-04-23 00:26:01 -0600

Having the same problem. eigen(Covariance, eigenval, eigenvec); //the resulting eigenvector is 10000by10000 matrix of all zeros (concerned, why are all zeros?) unable to deduce.

edit flag offensive delete link more

Question Tools

1 follower


Asked: 2017-03-11 10:19:47 -0600

Seen: 237 times

Last updated: Mar 11 '17