Three-dimensional visualization of multi-channel volume data: the amSFP algorithm

In this paper we present a three-dimensional visualization technique for multi-channel volume data. The technique simulates the physical process of fluorescence, hence its name: achromatic multi-channel simulated fluorescent process (amSFP). The data set is simulated as 3D distribution of different fluorescent dyes, where each channel is represented by a particular type of dye. Apart from the spatial density map, no additional characteristics about the data set have to be defined; no image segmentation is needed prior to visualization. The degree of interaction among the channels in the fluorescence process can be adapted to optimally render specific structures in the image. 3D multi-channel data can be obtained by a three-dimensional imaging device that is able to measure a number of physical quantities at a given location within a specimen. The fluorescence principle, the algorithm, and its implementation are presented. We have used the technique to investigate the relative spatial arrangement of blood vessels and astrocytes in the cat retina. The two components have been stained with different fluorescence dyes and recorded in a confocal light microscope to form a two-channel 3D data set.