Discrete Cosine Transform (DCT) is mainly used for image JPEG lossy compression. MIL can perform DCT using MimTransform(). For a one dimensional signal, this separates the signal into a set of cosine waves of different frequency. For a two-dimensional signal (image), it can be interpreted as the decomposition of an image into a set of 2D cosine patterns. The composition of these waves make up the original waveform. The forward DCT is defined as (for an 8x8 matrix):
where u and v are coordinates in the frequency domain.
The reverse DCT is defined as:
Frequency 0, also called the DC component, is plotted in the top-left corner of the spectrum. All other components in the spectrum are called AC components. A DCT concentrates the low frequency components of the image in the first few coefficients (top left-corner) of the spectrum. MIL divides the image into independent blocks of 8x8 pixels and performs the transform on each individual block. Centering of the spectrum is not supported in MIL.