The Classification of Laser Jet Typefaces by using Geometric Morphometries Technique and Principal Component Anal
Typeface classification is a fundamental issue in forensic document analysis. The main objective of typefaces analysis in forensic field is to determine what the precise typefaces that are being used during the typesetting time. There are many techniques that have already been published to classify...
Saved in:
Main Author: | |
---|---|
Format: | Monograph |
Language: | English |
Published: |
Universiti Sains Malaysia
2015
|
Subjects: | |
Online Access: | http://eprints.usm.my/60654/1/THONG%20WAI%20SENG%20-%20e.pdf http://eprints.usm.my/60654/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Typeface classification is a fundamental issue in forensic document analysis. The main
objective of typefaces analysis in forensic field is to determine what the precise typefaces that are being used during the typesetting time. There are many techniques that have already been published to classify typefaces. Those techniques include global texture analysis, typographical
features analysis, high-order statistical texture analysis and fractal multi-dimension analysis. In
this work, a new method of typefaces classification is proposed. This study employed geometric
morphometries technique coupled with the principle component analysis (PCA). Geometric
morphometries is a field of multivariate statistical analysis which is able to quantify shape and
size components of morphological variations. The purpose of this study is to classify 44
typefaces as well as to assess shape variation among typefaces of capital and lowercase k. The
results from this study indicate that geometric morphometries technique and principal component analysis can be employed to analysed shape variation of letter k between 44 typefaces as well as to classify typefaces. Some typefaces can be classified as one group as long as they have the
similar position of majority of their landmarks which contribute to their shape. |
---|