Wavelet Enhancement of Chromosome Banding Patterns

We will develop and commercialize significantly improved software for digital enhancement of the dechromosome banding patterns in microscopic images. Recently developed wavelet transform and multiresanalysis techniques, used with modern computerized chromosome analysis, promise improved enhancementdisplay of chromosome banding patterns, for more effective visual comparison of chromosomes to normaprototypes and to each other. This will help clinicians and researchers detect previously sub-visiblin conventional and high resolution banding. It will significantly increase the ability of automatedinstruments to assist the evaluation of chromosome alterations in clinical samples and in normal andcells. It was image enhancement by convolution filtering, coupled with high quality digital image recomputerized karyotyping first become practical almost a decade ago. Now multiresolution image proceconsiderably more potential than linear filtering for chromosome image enhancement, to make the mostlimited light microscope image. The recently-developed wavelet transform theory, and our early resulsuggest that these techniques offer much more enhancement potential than conventional convolution. Suse the biorthogonal wavelet transform or one of several overcomplete wavelet transforms to enhanceto bring out the relevant detail in the banding pattern while discriminating against components reflartifact, inter-chromosome variation, and other nongenetic differences. We will develop wavelet trancan, under operator control, selectively enhance image detail of different size and orientation, andwill implement this new capability on PSII's PowerGene commercial automatic karyotyping workstation,systems and as an upgrade to existing units, thereby making it broadly available to the cytogenetics