Multiresolution techniques for ultrasonic imaging

Multiresolution representations provide a hierarchical framework for analyzing the information content of images. At coarse resolution the decomposition provides global information about the image, whereas at finer resolution the details correspond to local features. This decomposition may be performed using the two dimensional wavelet transform with a quadrature minor filter bank architecture. In this paper, the techniques are applied to in vivo liver and breast sonograms to assess their ability to enhance the contrast between normal tissue and the abnormality of interest. The results are compared to nonlinear frequency diversity algorithms