Annotation Tool and XML Representation for Online Indic Data

In this paper we describe the semi-automatic annotation tool for annotating online handwritten data of Indic scripts. The annotation of handwriting data is essential to train and test the recognizers. In this paper we briefly describe the XML representation for storing online handwritten data in Indian languages. We then describe the annotation tool which essentially annotates at stroke, character and word level and exploits the uniqueness of XML standard to provide quality labels at different levels of annotation. The tool also facilitates classification of data based on quality of handwriting, age & region of writers. The annotator can verify the outputs suggested by the tool. The tool is supplemented by a utility for data segregation and accuracy calculator which aids quick performance analysis of recognizer. This tool is extensively used for annotating large amount of Hindi data and promising time saving is obtained in otherwise tedious annotation activity.