The WV_DWT function returns the multi-dimensional discrete wavelet transform of the input Array. The transform is done by WV_PWT using a user-inputted wavelet filter.
The length of each dimension of Array must be either a power of two (2), or must be less than four (4). The transform is not computed over dimensions of lengths less than four (4), but is computed over all other dimensions (for example, the wavelet transform of an array of size [3, 256] is computed over each [1, 256] column vector).
WV_DWT is based on the routine wtn described in section 13.10 of Numerical Recipes in C: The Art of Scientific Computing, 2nd ed. (Cambridge University Press), and is used by permission.
| Note |
Result = WV_DWT(Array, Scaling, Wavelet, Ioff, Joff [, /DOUBLE] [, /INVERSE])
The result is an output array of the same dimensions as Array, containing the discrete wavelet transform over each dimension.
The input vector or array. The length of each dimension must be either less than four (4) or a power of two (2).
A vector of scaling (father) coefficients, of length N.
A vector of wavelet (mother) coefficients, of length N.
An integer that specifies the support offset for Scaling. To center the scaling function over each point in Array, set Ioff to -N/2+2.
An integer that specifies the support offset for Wavelet. To center the wavelet function over each point in Array, set Joff to -N/2+2.
Set this keyword to force the computation to be done in double-precision arithmetic.
If set, the inverse transform is computed. By default, the forward transform is computed.
The WV_DWT function computes the wavelet coefficients using the pyramidal algorithm (Mallat 1989).
For a one-dimensional vector, the pyramid appears below:
Array elements [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15] \ / \ / \ / \ / \ / \ / \ / \ / s0,d0 s1,d1 s2,d2 s3,d3 s4,d4 s5,d5 s6,d6 s7,d7 \ / \ / \ / \ / \ / \ / \ / \ / S0,D0 S1,D1 S2,D2 S3,D3 \ / \ / \ / \ / \ / \ / S0,D0 S1,D1
At each level of the hierarchy, the WV_PWT function is used to compute the scaling coefficient Si and wavelet coefficient Di (where i represents the position). The letters s, S, S and d, D, D represent increasing scale. The wavelet coefficients are stored in Result in order from largest scales to smallest:
Result = [ S0, S1, D0, D1, D0, D1, D2, D3, d0, d1, d2, d3, d4, d5, d6, d7 ]
For a two-dimensional Array, the wavelet transform is computed using the pyramidal algorithm along each dimension. The wavelet coefficients are stored in order with the largest scales in the [0, 0] position. As an example, for an 8 x 8 Array, the Result is an 8 x 8 array with the following structure:
[0,0] [[ A0B0 A1B0 C0B0 C1B0 c0B0 c1B0 c2B0 c3B0 ], [ A0B1 A1B1 C0B1 C1B1 c0B1 c1B1 c2B1 c3B1 ], [ A0D0 A1D0 C0D0 C1D0 c0D0 c1D0 c2D0 c3D0 ], [ A0D1 A1D1 C0D1 C1D1 c0D1 c1D1 c2D1 c3D1 ], [ A0d0 A1d0 C0d0 C1d0 c0d0 c1d0 c2d0 c3d0 ], [ A0d1 A1d1 C0d1 C1d1 c0d1 c1d1 c2d1 c3d1 ], [ A0d2 A1d2 C0d2 C1d2 c0d2 c1d2 c2d2 c3d2 ], [ A0d3 A1d3 C0d3 C1d3 c0d3 c1d3 c2d3 c3d3 ]]
Here A and B represent the scale coefficients for the first and second dimensions, respectively, The C and D represent the largest-scale wavelet coefficients for the first and second dimensions, respectively, while c and d represent the small-scale wavelet coefficients. Subscripts 0, 1, 2, 3 denote the position of the wavelet within the image.
Introduced: 5.3