Transc&Anno: A Graphical Tool for the Transcription and On-the-Fly Annotation of Handwritten Documents

Transc&Anno: A Graphical Tool for the Transcription and On-the-Fly Annotation of Handwritten Documents Nadezda Okinina, Lionel Nicolas, Verena Lyding Eurac Research, viale Druso 1, 39100 Bolzano, Italy nadezda.okinina@eurac.edu, lionel.nicolas@eurac.edu, verena.lyding@eurac.edu Abstract We present Transc&Anno, a web-based collaboration tool allowing the transcription of text images and their shallow on-the-fly annotation. Transc&Anno was originally developed in order to address the needs of learner corpora research so as to facilitate digitisation of handwritten learner essays. However, the tool can be used for the creation of any type of corpora requiring transcription and shallow on-the-fly annotation resulting in inline XML. Transc&Anno provides an intuitive environment that is explicitly designed to facilitate the transcription and annotation process for linguists. Transc&Anno ensures a high transcription output quality by validating the XML and only allowing predefined tags. It was created on top of the FromThePage transcription tool developed entirely with standard web technologies – Ruby on Rails, Javascript, HTML, and CSS. We adapted this open-source web-based tool to linguistic research purposes by adding linguistic annotation functionalities to it. Thereby we united the convenience of a collaborative transcription tool with its advanced image visualisation, centralised data storage, version control and inter-collaborator communication facilities with the precision of a linguistic annotation tool with its well-developed tag definition possibilities, easy tagging process and tagged-text visualisation. Transc&Anno is easily customisable, open source, and available on Github. Keywords: transcription tools, annotation tools, learner corpora 1. Introduction Linguistic resources are of fundamental importance for conducting linguistic research. Some of these resources are originally created in handwritten form but need to be digitised in order to be efficiently exploited. In the absence of automatic handwriting recognition systems capable of handling in a satisfactory fashion the inputted data, they have to be manually transcribed. In many cases, transcription can coincide with shallow annotation. Since the transcription and annotation process can be extremely laborious and time-consuming, Transc&Anno is designed to make transcription and annotation quick and intuitive. Transc&Anno was specifically developed in order to address the needs of learner corpora research (Glaznieks et al., 2014) with regard to the manual transcription of learner texts and their shallow on-the-fly annotation. In the remainder of this paper, we present the Learner Corpus Infrastructure project that led to the creation of Transc&Anno (Section 2), review existing text transcription and annotation tools (Section 3), explain Transc&Anno’s use and important features (Section 4), give some technical information (Section 5), point out future development possibilities and conclude the paper (Section 6). 2. The LCI Project 2.1 Overview The Learner Corpus Infrastructure project (LCI) conducted at the Institute for Applied Linguistics at Eurac Research aims at creating a stable, systematic, and sustainable infrastructure for the collection, processing and maintenance of learner corpora (Nicolas et al., 2015). “Computer learner corpora are electronic collections of authentic FL1/SL2 textual data assembled according to explicit design criteria for a particular SLA3/FLT4 purpose. They are encoded in a standardised and 1homogeneous way and documented as to their origin and foreign language 2 second language 3 second language acquisition …

Content truncated to 500 words