Handwriting Analysis and Synthesis

ISBN-13: 9783838363998 / Angielski / Miękka / 2010 / 124 str.

Ashish Gupta

Opis

The objective of this thesis is to learn the handwriting of a person and artificially synthesize text that is indistinguishable from that persons handwriting. The thesis builds this system for writer specific artificial handwriting synthesis from scratch. It builds a handwriting sampling text which aims at acquiring maximum individualistic handwriting traits; minimizing sampling text word count; retaining fidelity of natural handwriting. It analyzes image processing techniques to reduce image acquisition noise and extract information relevant to modelling handwriting. It shows how the handwritten stroke and letter shape is modelled. It demonstrates how writer specific realistic handwriting is synthesized by pseudo-random variations in the parameters of the model of that writers handwriting. Finally, it indicates various avenues for further research and development to scale the system to higher complexity to synthesize more realistic handwriting.

O autorze

Ashish Gupta is a PhD student at the University of Surrey. Previously, he has worked at Research Engineer at Sarnoff Corporation for two years. He holds a Master s degree in Electrical Engineering from the prestigious Indian Institute of Technology. His current research interests are Computer Vision, Pattern Recognition, and Machine Learning.