原书出处
Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)
For stationary signals, the covariance matrix C has a circulant structure (assuming a tiling of the image) and can be diagonalized using the Fourier transform. The eigenvalues of the diagonal matrix correspond to the power (squared magnitude) of the frequency components of the Fourier transform. This is, the matrix C can be written as C = FDFT. Where F is the matrix that contains the
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c027b.xhtml
EPUB SHA-256:60a7eebc43f5d9a675d776fed3736fbdaa791e55877023f2f8cba016a9b384f8