Language-independent speech to text transcript matching

Abstract

The main business targets of Glavitsch Eggler Software are language learning and retrieval/indexing of audio archives. Readylingua.com offers audio recordings aligned with transcripts for individuals and language schools. Today, the audio recordings and transcripts have to be aligned manually with a software called SpeechIndexer. This manual aligning is very time consuming and limits the activities of the startup company. The goal of this project is to develop algorithms supporting the alignment process for multiple languages and for partially inexact transcripts. These algorithms will be integrated into SpeechIndexer, opening the additional possibility to license SpeechIndexer to language centers, schools like Klubschule Migros and companies or institutions that maintain archives with audio/video recordings, e.g. parliamentary services and radio/TV stations.