An attempt for coloring multichannel MR imaging data

This is an elementary research for assigning color values to voxels of mult ichannel Magnetic Resonance Imaging (MRI) volume data. The MRI volume data sets obtained under different scanning conditions are transformed to the co mponents by independent component analysis (ICA), which enhances physical c haracteristics of the tissue. The transfer functions for generating color v alues from independent components are obtained using the radial basis funct ion network, a kind of neural net, by training the network with sample data chosen from visible human female data set (VHF). The resultant color volum e data sets correspond well with the full-color cross-sections of the visib le human data sets.