Gaussian Variogram Model for Printing Technology Identification

Abstract

Tampering of documents is monotonically growing by posing challenges to forensic scientists. There is a great need to develop alternative solutions for forensic characterization of printers. This paper analyzes documents printed by various printers and characterizes them for identification purposes. Present study focuses on developing a model Gaussian Variogram Model (GVM) for identifying the print technology which produced the given document. This method characterizes print technology based on spatial variability. Homogeneous color region of images are taken as samples for the GVM data generation. The generated GVM data is taken as input to generate Reduct based Decision Tree (RDT), which gives rules to identify the source printer for the given test data. Performance analysis of the model is also presented. Developed method assists the document examiner in finding basic print pattern of printers and it is also helpful in classifying different print technology.