Karhunen–Loève functions represent the best choice for modal wavefront reconstruction. They are usually built up as a linear combination of Zernike polynomials by using principal component analysis methods; thus they are ordered by covariance. Using Shannon information theory, we provide a best reordering procedure based on the concept of mutual information. This enhances reconstruction efficiency, allowing us to reduce the basis dimension while maintaining the same fitting error in wavefront reconstruction.
© 2010 Optical Society of America
Equations on this page are rendered with MathJax. Learn more.