/r/-Letter disorder diagnosis (/r/-LDD): Arabic speech database development for automatic diagnosis of childhood speech disorders (Case study)

摘要

In light of the scarcity of both published and free Acoustic Arabic databases, we propose in this paper Acoustic Arabic database to be a reference in the field of automatic Arabic speech recognition, this database is the result of a case study that has been developed to contribute to the automatic diagnosis of speech disorders in Arabic speaking children, the field work was in collaboration with experts in communication and relying on some multinational Arabic schools to record samples of various Arabic speech dialects in normal circumstances. The letter "R" has been selected as the most common letters that children suffer from. In this paper we will explain the mechanism of the development and design of this database, which is divided into three sub databases: the first is for diagnosis of the disorder in a letter "R" when its position is in the beginning of the word and the second when it is in the middle of the word the last one when it is at end of the word, each sub database contains speech recordings for 60 children; 30 males and 30 females, each child repeats the voice disorder five times. We hope that this acoustic database will be considered challenge for working more and to be a reference for future researches as the identification of Arabic speech in general and especially for the automatic diagnosis and treatment of speech disorders for children.