Abstract
Handwritten documents can be characterized by their content or by the shape of the written characters. We focus on the problem of comparing a person's handwriting to a document of unknown provenance using the shape of the writing, as is done in forensic applications. To do so, we first propose a method for processing scanned handwritten documents to decompose the writing into small graphical structures, often corresponding to letters. We then introduce a measure of distance between two such structures that is inspired by the graph edit distance, and a measure of center for a collection of the graphs. These measurements are the basis for an outlier tolerant K‐means algorithm to cluster the graphs based on structural attributes, thus creating a template for sorting new documents. Finally, we present a Bayesian hierarchical model to capture the propensity of a writer for producing graphs that are assigned to certain clusters. We illustrate the methods using documents from the Computer Vision Lab dataset. We show results of the identification task under the cluster assignments and compare to the same modeling, but with a less flexible grouping method that is not tolerant of incidental strokes or outliers.
Highlights
Many disciplines rely on the ability to parse, process, and analyze handwritten text
The groupings that result from the dynamical clustering method we propose are more parsimonious, descriptive, and repeatable for writers than deterministic groupings, because of their robustness to small structural differences among graphs
It is worth mentioning that we only use handwriter for document processing in this work, but the software has other feature extraction capabilities such as finding centroids, slants, loops, and other measurable attributes for each graph. These features are undeniably important for forensic handwriting analysis, but are not used to create the clustering template that is of focus here, and we do not discuss them further
Summary
Many disciplines rely on the ability to parse, process, and analyze handwritten text. We use scanned handwritten documents from a variety of writers in the Computer Vision Lab (CVL) database [8] to meet our goals In this writership analysis framework, there are two stages of feature extraction that occur. A question of interest is whether writers can be distinguished by the proportion of the graphs extracted from their writing that fall into each of the k clusters in the template To address this question, we use the observed cluster frequencies in a document by a writer as the response variable in a hierarchical model to estimate the posterior probability of writership for each writer in a closed set. The groupings that result from the dynamical clustering method we propose are more parsimonious, descriptive, and repeatable for writers than deterministic groupings, because of their robustness to small structural differences among graphs.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.