ABSTRACT
This chapter reviews wavelets, discrete cosine transform, Gabor filters, Zernike filters, Fourier filters, fractional power filters, pulse-coupled neural networks and empirical mode decomposition. The discrete cosine transform (DCT) is sensitive to frequencies within the image. Multiple cosine functions are applied, thus producing several outputs that depict the strength of each frequency inherent in image. Gabor filters are sensitive to edges and orientation within an image. This type of filters considers the same information as does the Wavelet transforms. The DCT responds to frequencies and orientations that are inherent in the image. The Zernike polynomials create a basis set but these are more sensitive to radial and polar changes. Empirical mode decomposition (EMD) is a method by which a signal is decomposed into a set of intrinsic mode functions which are sensitive to different frequency bands of the original signal. There are some concerns that accompany the EMD process, which need to be addressed before it can become an automated tool.
