Finger motion in classifying offline handwriting patterns

Loading...
Thumbnail Image
Date
2017-05
Authors
Yeoh, Shen Horng
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Offline handwriting recognition refers to the ability of a machine to receive and interpret a previous individual-made handwritten input from a photographed or scanned image. In previous studies, the offline handwriting classification is determined solely based on the handwriting patterns. To the best of our knowledge, no studies were found to predict the English words inclination based on the finger motions. Therefore, this study aims to relate the finger movements to handwriting patterns. The specific objectives include: (i) to determine whether finger motion attributes can distinguish patterns of handwriting, (ii) classify handwriting patterns by sentence inclination based on different finger motion, (iii) to investigate the rule-reasoning statements between the finger motion and the handwriting inclinations. This study involves the features extractions from handwriting patterns of 30 subjects with recorded videos of finger movements during writings. Raw data undergo three stages of data mining analyses; data preprocessing, data classification and data interpretation. The preprocessed data is classified using the J48 tree algorithm. The correctly classified accuracy prediction after trained could achieve up to 98 %, Finding revealed that the angle of thumbs plays a significant role in classification of the inclination of the English sentence.
Description
Keywords
Citation