E!11562 e-Lexia STTS Södermalms Talteknologiservice AB

Purpose and goal

The overall goal of the project was to create a learning platform for dyslectic children, including phoneme based recognition to ready for commercialization after project ending. We find that the project has been successful in that the result of the project is a learning platform that is ready to be launched and that we have created a promising prototype with the speech recognition included. Before releasing a version with ASR included, further data collections and tests need to be made. STTS has developed software and methods to be able to develop and offer ASR.

Expected results and effects

For STTS the project has given us the opportunity to gain knowledge and to develop and to collect our own software and methods. This will, and has to a large extent already, made the company more attractive on the market both as a subcontractor and partner. Many of the tools and methods will be useful also for general ASR which opens up many opportunities which we are exploring at the moment. It is of course also our aim and hope to be able to take the phoneme based recognition to a commercial level for the e-Lexia platform which would create great opportunities for generating revenue.

Planned approach and implementation

The e-Lexia project had 6 work packages. The consortium has in general followed this structure and we have a result for all work packages, although for some the scope had to be reduced. The main challenge for the consortium has been to ensure compliance with the ethical requirements made by Eurostars. Difficulties to correctly interpret and secure compliance delayed any work with children and the data collections could not be made as originally planned. STTS has had to settle for training with smaller sets of data and additional data collections are needed for the ASR to be ready for launch.