I am using PCA for dimension reduction .After some preprocessing i have stored my images in a matrix . In matrix i found that some of pixel value coming like NaN+NaNi . How to overcome with that so that covariance will b calculated .

A is the matrix where all images are stored column wise in which NaN+NaNi is there . cov=A' * A. covariance matrix is full of NaN+NaNi so, further processing is not possible to find eigen values and eigen vector. Please reply soon i stucked on this step on my project work . Thank you