Browse Prior Art Database

Human Variation Font System

IP.com Disclosure Number: IPCOM000020614D
Original Publication Date: 2003-Dec-03
Included in the Prior Art Database: 2003-Dec-03
Document File: 2 page(s) / 10K

Publishing Venue

IBM

Abstract

Disclosed is software for a rule based font system. The rules determine which variety of a particular character will be displayed. The user may reconfigure the rules.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 1 of 2

Human Variation Font System

Disclosed is software for a rule based font system. The rules determine which variety of a particular character will be displayed. The user may reconfigure the rules.

Handwriting fonts are licensed and/or purchased by customers including advertisers, direct mailing companies, application developers, and end users. Handwriting fonts are intended to simulate actual human handwriting. In the United States, print and cursive font sets are common. These fonts do not account for human variation. They are too perfect. For example, all the lower case "e"s are perfectly identical. Ultimately, the human eye knows the handwriting fonts are not actual human handwriting. As a result, in some cases, the marketing value or customer appeal is lost when the reader realizes they are reading machine printed words.

The disclosed software is a handwriting font system that accounts for human variation. This system has configurable rules that decide which character variation is appropriate. When trained sufficiently, the handwriting font system will be indistinguishable from the font author's actual handwriting. The disclosed handwriting font system accounts for:
1) character combinations; standalone characters and around other characters)
2) location; mimics writer fatigue between beginning and end of a document)
3) the ability to match instances directly or select them in a sequential or randomized fashion.

The disclosed handwriting font system must be trained sufficiently before it will work properly.

An example capture implementation may include the following steps:

Step 1: Scan handwritten document using a high resolution flatbed scanner. At a minimum, this document must include the characters and words defined on the font training template. The template includes writing lowercase and uppercase alphabetic characters, numeric characters, punctuation and symbol characters, and words with frequently used character combinations. Each item on the font training template is written five times each. Additional documents may be scanned to increase human variation accuracy.

Step 2: Capture character variations via OCR, or similar technology, conversion. Example: "The person connected to the ethernet!"
a) The OCR engine recognizes "The " as a "T" followed by "h" followed by "e" followed by " ".
b) The handwriting agent catalogues the scanned "T" as this is what a "T" looks like at the start of a sentence or paragraph followed by an "h". The agent catalogues the scanned "h" as this is what an "h" looks like between a "T" and an "e". The agent catalogues the scanned "e" as this is what an "e" looks like after an "h" and as the final character in a word. Each character is catalogued along with its line number in the document.
c)...