Lipi-PocketPCDemo: A Sample Windows Mobile Application for Isolated Character Recognition

Overview

This is a sample application that uses Lipi Core Toolkit 2.2 for recognizing English lowercase alphabets on Windows Mobile Pocket PC. This application can be used for testing user defined pre-build recognizers. The application source code as can be used as a reference to build handwriting recognition applications.

Supported Platform

Windows Mobile 6.0

Installation instructions

Binary package: lipi-PocketPCDemo1.0-wm.cab//
► Use the Cabarc SDK (Refer Appendix for download instructions) tool to extract the lipi-'PocketPCDemo'1.0-wm.cab file in Windows PC. The package will be extracted to the lipi-'PocketPCDemo'1.0-wm folder.
► Use ActiveSync to transfer the lipi-'PocketPCDemo'1.0-wm/bin/lipi-'PocketPCDemo'.CAB from the Windows PC to the Pocket PC device.
► Locate the lipi-'PocketPCDemo'.CAB file in the Pocket PC (default location is 'My Documents') and then tap on the file to install the application.
► The sample application packaged with Lipi Core Toolkit 2.2 library and pre-built recognizer for English lowercase alphabets will be deployed.
Note: You may refer the PocetPC Central Tutorial for information on installing a CAB file on Windows Mobile.

User interface

A snapshot of the User Interface is given below.

Launch the application from the installed location (Open 'File explorer' and locate 'LIPITK' folder) in Windows Mobile. On start up, the application loads the Lipi Core Toolkit library and instantiates the Lipi recognition engine.
Load Project: Click on the Load --> Project menu item button to load the English lowercase pre-built recognizer.
Note: The LIPITK/projects directory must have the pre-built recognizer data files.
Writing Area: Write a lowercase English alphabet in the writing area. As soon as the user completes writing, the top 4 choices of recognized character are displayed in the choices panel.
Clear: The Clear Menu Item can be used to clear the results displayed in the choices panel.

Changing the default pre-built recognizer

To use an alternative pre-built recognizer instead of the default English lowercase data, replace the files under LIPITK/projects directory with the pre-built recognizer of choice.

Preparing Unicode Mapping file

The application uses the file LIPITK/projects'/<pre-built recognizer>/config/mapping.txt// to map the class IDs returned by the recognizer to the corresponding character Unicode. This is then displayed in the choices panel. Users may refer Unicode Charts for finding the Unicode character code by scripts. If this mapping file relevant for the user defined pre-built recognizer is not provided, then the application displays the Class IDs received from the recognizer in the choices panel.