Experimental Handwriting System

Pace CSIS is conducting a study to validate the individuality of handwriting using a characteristic set of handwriting features and to discover knowledge useful to forensic document examiners by performing statistical hypothesis testing. The work is expected to have a bearing on the admissibility of handwriting testimony in US courts.

The project has three parts: image database construction, feature extraction (image processing), and machine learning and data mining. Students will develop the interface and the platform for the handwriting analysis system. Various machine learning algorithms like k-means clustering, neural network, nearest neighbor, etc., will be explored. Data mining techniques, such as the apriori algorithm, will also be studied.

Since this is a rather large project, the work of the student team might be limited to several of the following sub-topics:

Note: If the project funding award is granted by the National Institute of Justice, students involved in this project will be candidates for graduate research assistantships.