PCA returns wrong eigenvector (Bug #4335)

#include <iostream>
#include "opencv2/core.hpp" 
#include "opencv2/ml.hpp" 

int main() {
    // ingredients matrix from hald data in MATLAB
    cv::Mat data = (cv::Mat_<float>(13, 4) <<\
            7, 26, 6, 60,\
            1, 29, 15, 52,\
            11, 56, 8, 20,\
            11, 31, 8, 47,\
            7, 52, 6, 33,\
            11, 55, 9, 22,\
            3, 71, 17, 6,\
            1, 31, 22, 44,\
            2, 54, 18, 22,\
            21, 47, 4, 26,\
            1, 40, 23, 34,\
            11, 66, 9, 12,\
            10, 68, 8, 12);

    cv::PCA pca(data, cv::Mat(), cv::PCA::DATA_AS_ROW);
    cv::Mat eigenvectors;
    cv::Mat mean;
    eigenvectors = pca.eigenvectors.t();
    mean = pca.mean;  // row vector
    std::cout << eigenvectors << std::endl;
//    std::cout << mean << std::endl;
//    std::cout << (data - cv::repeat(mean, 13, 1)) * eigenvectors.t() << std::endl;

    return 0;
}

The correct eigenvector matrix is
-0.0678 -0.6460 0.5673 0.5062
-0.6785 -0.0200 -0.5440 0.4933
0.0290 0.7553 0.4036 0.5156
0.7309 -0.1085 -0.4684 0.4844

but pca returns
-0.0678 -0.6460 -0.5673 0.5062
-0.6785 -0.0200 0.5440 0.4933
0.0290 0.7553 -0.4036 0.5156
0.7309 -0.1085 0.4684 0.4844

The third column has opposite sign.