Assisting Forensic Writer Verification by Visualizing Diversity of Digit Handwritings: An Approach by Multidimensional Scaling of Earth Mover's Distance

In this paper, we demonstrated a method to assist forensic handwriting experts in the process of writer verification by visualizing the diversity of overall shapes of digit handwritings. Multidimensional scaling (MDS) was applied to Earth Mover's Distance (EMD) data calculated between 60 different writers in order to visualize each writer's feature in population. EMD flow was calculated between k-means cluster centroids, which represented kernel density distribution of handwritten stroke of each writer within six trials. Experimental results for all ten kinds of digit handwritings showed that the relation of overall handwritten shapes between writers was successfully visualized as the locus in multidimensional space. The meanings given to each MDS axis, such as "slant", "width" and "the balance between parts", were almost common through all kinds of digit. They were corresponded to the feature which have been paid attention by forensic handwriting experts and pointed out in previous studies or literatures. The state of distribution such as the density in multidimensional space was considered to provide effective information to forensic examiners in evaluating the rarity of handwritten features observed in questioned document.