Texture Traits with Uniform-Quantization in Handwriting Documents for Digital Forensics Investigation
Dian Pratiwi, Syaifudin, TRUBUS RAHARDIANSYAH, Anastasia Jocelyn Hilman, Wendelina Anggriani, Nikita Chairunnisa · IOP Conference Series Materials Science and Engineering · 2019
Abstract In crime and falsification related to documents are often difficult to verify orauthenticated by the authorities. This is the basis of research to develop a system in assisting digital forensics to investigate and seek the truth of the evidence in the form of digital handwritten documents. The steps that being taken by researchers are collecting and digitizing the documents into image form, converting color from RGB to greyscale, separate the object through thresholding, color histogram, uniform quantization from 256 to 128 of greylevel, texture feature extraction ie variance, skew, relative smoothness, entropy and mean, normalize the feature value and similarity measures through Euclidean distance calculations. From the results of testing of 10 data that has been matched with 20 training data, 6 documents successfully recognized correctly the authenticity of the document owner. Thus, the system in this study produces an accuracy of 60% and can be used to assist digital forensics in analyzing the authenticity of handwritten documents.