Abstract

When a person composes a document by hand, there is random variability in what is produced. That is, every letter is different from all others. If the person produces seven a s, none will be the same. This is not true when a computer prints something. When the computer produces seven a s they are all exactly the same. However, even with the variability inherent in a person s handwriting, when two people write something and they are compared side by side, they often appear as different as fonts from two computer families. In fact, if the two were intermixed to produce some text that has characters from each hand, it would not look right! The goal of this application is to improve the ability to digitally create testing materials (i. e., data collection documents) that give the appearance of being filled out manually (that is, by a person). We developed a set of capabilities that allow us to generate digital test decks using a raster database of handprinted characters, organized into hands (a single person s handprint). We wish to expand these capabilities using vector characters. The raster database has much utility to produce digital test deck materials. Vector characters, it is hoped, will allow greater control to morph the digital test data, within certain constraints. The long-term goal is to have a valid set of computer-generated hands that is virtually indistinguishable from characters created by a person.

Publication Date

2006

Document Type

Master's Project

Student Type

Graduate

Advisor

Misic, Vladimir - Chair

Advisor/Committee Member

Anderson, Peter

Advisor/Committee Member

Paxton, K. Bradley

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works in December 2013.

Campus

RIT – Main Campus

Download

Share

COinS