Ethiopic Online Handwriting Recognition in Android-Based Smartphones
No Thumbnail Available
Date
2010-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
A rapid growth in advanced mobile technologies has opened various research interests.
Handheld devices like Smartphones are getting a wide range of acceptance due to the fullfledged
functionalities they provide. They are being designed to have the ability to access
information of any type ubiquitously.
Text entry mechanisms are main concerns in handheld devices as they are the core interaction
ways of communication. To best use of all features incorporated in these devices, an efficient
text entry mechanism is vital. Considering the aim to achieve the goals of being small in size
to hold and easy to use, made to look for alternative text entry techniques rather than
sticking to the hardware keyboard which resulted in difficulties to attain the goals set.
Using the touch screen capabilities of handhelds, online handwriting recognition and soft
keyboard entries are becoming better options for text entry into these devices considering the
inconvenience of hardware keyboards to be used.
In this work, an Ethiopic Online Handwriting Recognition (EOHWR) input method is
developed for an Android-based Smartphones where users of the script shall benefit from it
by using the developed input method in different applications that exist on the device. In its
integration, the work considered previous efforts exerted on the EOHWR.
The work followed a constrained, writer independent approach for its recognition. It
developed representation symbols for thirty four basic characters, a numeral and punctuation
of the Ethiopic script resembling the natural script and stored their references persistently.
Users need to learn the symbols prior to usage. It also puts some restrictions on the orders in
which strokes that build a character should be inserted. In its recognition, it followed a
template matching technique computing the directional distances between segments of a
stroke for each character.
The work incorporates the usage of the non basic characters, numerals and punctuations
which are found in the Ethiopic script by making them to be candidate soft key options of
their corresponding basic symbol. Thus, whenever recognition takes place for a basic symbol,
its non basic character family will be available to be selected. Its usability and accuracy is
affirmed getting an average accuracy of about 80 percent.
Keywords: Ethiopic Input Method for Android, Constrained Ethiopic Online Handwriting
Recognition System, Writer Independent Ethiopic Online Handwriting Recognition System
Description
Keywords
Ethiopic Input Method for Android; Constrained Ethiopic Online Handwriting Recognition System; Writer Independent Ethiopic Online Handwriting Recognition System