Cfwvq Compresses a color image by vector quantizing its wavelet transform
Cfwvq [-R Rate] [-c] [-m] [-s] [-u] [-r NLevel] [-b1 BiFilt1] [-b2 BiFilt2] [-o OrthoFilt] [-e EdgeFilt] image codebook_red [codebook_green codebook_blue]
This macro compresses a RGB color image using a vector quantization algorithm applied to the orthogonal/biorthogonal wavelet coefficients. It is based on the fwvq module (which is applied separately to each channel), and is simpler to use.
It works exactly in the same way that the Fwvq macro for graylevel images, except that there is a possibility to use different codebooks for the red, green and blue channels. If the optional argument codebook_green and codebook_blue are activated, then the codebook sets whose prefix is codebook_red are used to compress the red channel, the codebook sets whose prefix is codebook_green are used to compress the green channel and the codebook sets whose prefix is codebook_blue are used to compress the blue channel. Otherwise the same codebook sets whose prefix is codebook_red are used to compress all three red, green and blue channels.
The target rate (if -R option is selected) is the same for red, green and blue channels. Notice that Rate is a target bit rate per point and per channel. Three separate compressed files are generated (*_Rater.comp, *_Rateg.comp and *_Rateb.comp if the input image file is *.rim or *.img). The quantized image is put in *_Rateq.rim.
Like in Fwvq the dimensions of the image should have a minimum of factor 2 in their decomposition in prime numbers. If not then a part of the image as large as possible is extracted and having the required number of factor 2.
Notice that codebook_red, codebook_green
and codebook_blue are prefixes for one channel codebook sets
which may be simply generated with Fwlbg_adap macro.
One must however first extract the channel images of the training color
images using the cfcolor2channels module.
Last Modification date : Feb 11 2002
Jean-Pierre D'Ales, Jacques Froment