Vibrating exposure on on-line and off-line handwriting recognition
Loading...
Date
2017-06
Authors
Wong, Liang Chern
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Handwriting recognition refers to the transformation of a language into symbolic representation from its visual marks [1]. Off-line handwriting recognition involves the interpretation and conversion of handwritten character from a hardcopy, whereas on-line handwriting recognition involves constant dynamic conversion of text as it is written on a digital device. Previous related work has focused on the techniques and methods applied in feature extraction. However, no study has reported on effects of vibration to the handwriting patterns. Handwritings with exposure to vibration tend to show variation from original handwriting style. Hence, this study applies the data mining technique to analyse the features extracted from the handwritings to compare between with and without vibration impact. The objectives of the study are to (i) study the effects of vibration to the recognition of the handwriting characteristic, (ii) classify handwriting by the vibration, and (iii) relate the differences in handwritten characters recognition by the vibrating impact. Four stages of data analyses are involved: data collection, pre-processing, processing, and knowledge discovery. KStar, PART, and J48 classification algorithms were used to classify the data into four predefined classes: on-line normal, on-line vibration, off-line normal, and off-line vibration. The percentage of classification accuracy achieved from the three algorithms are 100%, 92%, and 96% respectively. Whether with or without vibration impact, case study findings revealed that the length of last word and second measurement of size are the crucial identities of a person’s handwriting.