Evaluation of genetic algorithm-generated multivariate color tables for the visualization of multimodal medical fused data sets

摘要

Application of a multimodality imaging approach is advantageous for detection, diagnosis, and management of many ailments. Display is limited to two or three dimensions when using spatial relationships alone. The use of color, in addition to spatial relationships increases the dimensionality of the data that can be effectively visualized. A genetic algorithm has been developed to automatically generate color tables satisfying defined requirements for the fused display of high-resolution and dynamic contrast-enhanced magnetic resonance imaging and F18-FDG positron emission tomography data sets. Radiologists were asked to evaluate images created using several different fusion-for-visualization techniques. The study determined radiologists’ preference, ease of use, understanding, efficiency, and accuracy when reading images using each technique. The genetic algorithm generated color tables were rated as the preferred ones.