Creating and sharing knowledge for telecommunications

Project: Automatic assessment of reading ability of children

Main Objective:
This work aims to develop a technological module for Portuguese children, able to detect mispronunciations, hesitations and other disfluency indicators, in order to compute an overall reading ability index. The main problem consists in the detection of mispronunciations, which requires detailed acoustic models and new detection techniques dealing with speech variability using models of allowable pronunciations. The developed methods can be put to use in a prototype application, useful not only for the children, but also for the teachers and tutors that are involved in their education. The proposed system could be used in the near future in primary schools as a complimentary way of assessment of students or as a speech therapy tool.
Reference: internal project
Funding: IT/LA
Start Date: 01-02-2014
End Date: 01-03-2016
Team: Fernando Manuel Santos Perdigao, Jorge Daniel Leonardo Proença, Carla Alexandra Calado Lopes, Sara Maria Fernandes Rato e Costa Marques Candeias, Arlindo Oliveira da Veiga, Miguel Sales Dias
Groups: Multimedia Signal Processing – Co, Multimedia Signal Processing – Lr
Partners: MLDC
Local Coordinator: Fernando Manuel Santos Perdigao
Associated Publications