An orthogonal wavelet representation of multivalued images

In this paper, a new orthogonal wavelet representation for multivalued images is presented. The idea for this representation is based on the determination of the maximal length of linear vector operators, applied to multivalued images. This concept is derived from the concept of maximal gradient of multivalued images. The pyramidal dyadic wavelet transform algorithm, using quadrature mirror filters is modified to be applied to multivalued images, where all detail information of the different bands is accumulated into one single representation. The representation allows for a multiscale edge description of multivalued images, with several applications ranging from multispectral image fusion to multivalued image enhancement, denoising and segmentation. 1