Weighted Distance Mapping (WDM)

A new method is introduced for describing, visualizing, and inspecting data spaces. It is based on an adapted version of the Fast Exact Euclidean Distance (FEED) transform. It computes a description of the complete data space based on partial data. Combined with a metric, a true Weighted Distance Map (WDM) can be computed, which can define a probability space. Subsequently, distances between data points can be determined. Using edge detection, borders (or boundaries) between categories (or clusters) of data can be found. Hence, Voronoi diagrams can be created. Moreover, the visualization of such WDMs provides excellent means for data inspection. Several examples illustrate the use of WDMs as well as their efficiency. So, a new, fast, and exact data analysis method has been developed that yields the means for a rich and intuitive method of data inspection.