Adaptive Pseudocoloring of Medical Images Using Dynamic Optimal Partitioning and Space-Filling Curves

Pseudocoloring is a popular technique that assigns colors to gray level images in order to capitalize on the human´s eyes better sensitivity to color changes over gray level intensity changes. The method is commonly used in many applications including visualization techniques, thermal imaging, quality inspection and art creation. In this paper, we present a new algorithm for pseudo coloring, which combines space-filling curve representation of the image with optimal partitioning with dynamic programming. The algorithm allows the assignment of triplet colors to the pixels of an image in a fine to coarse fashion, thus creating either a detail-preserving or structure preserving pseudo-colored image. This is achieved by starting with the original gray levels present in the image and reducing them with optimal partitioning down to any specified number (including two colors for a binarized version of the original image if desired). The Algorithm outputs the number of colors, and the actual assignment of colors follows depending on the problem at hand. Application of the algorithm is demonstrated for a set of two medical images.