Avaliação de um programa informatizado de reconhecimento de fala em indivíduos com deficiência mental e com problemas de linguagem.

• Ano de defesa: 2005
• Autor(a) principal: Carrer, Henildes José
• Orientador(a): Goyos, Antônio Celso de Noronha
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Educação Especial - PPGEEs
• País: BR
• Palavras-chave em Português: Estudantes com necessidades educacionais especiais; Reconhecimento de fala; Ensino informatizado
• Palavras-chave em Inglês: Speech recognition; Computer teach; Phonologic disturb
• Área do conhecimento CNPq: CIENCIAS HUMANAS::EDUCACAO::TOPICOS ESPECIFICOS DE EDUCACAO::EDUCACAO ESPECIAL
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/3198

The general aim of this study was the evaluation of the speech-recognition technology together with the MESTRE software, in order to check their viability to educational objectives. 110 individuals participated of this study, divided in four groups: G1, constituted by 10 university students; G2, constituted by 60 children presenting normal development: 20 children from 4 years to 4 years and 11 months; 20 children from 5 years to 6 years and 11 months and 20 children from 7 to 7 years and 11 months; G3, constituted by 20 individuals with diagnosis of mental deficiency, from 7 to 14 year old, and G4, constituted by 20 individuals from 7 to 11 year old with diagnosis of phonologic disturbs. The experimental stimuli were presented through the software MESTRE, constituted by two sets, A and B. The A set formed by 51 words of the Portuguese language and the B set was formed by 51 pictures related to the words of the A set. The tested relations were: spoken word - oral production (AE relations); picture oral production (BE relations). The results were statistically analyzed. The average of right answers of the AE relation tasks was compared to the BE relation of the same group of stimuli and the differences were considered non-significant. In the analysis by groups of individuals, the average of right answers in each tested relation was: in G1, in the AE relation, 90,51% and in BE relation, 90,49%; in G2, from 4 years to 4 years and 11 months, the average was 55,4% in AE relation and 52,4% in BE relation; from 5 years to 6 years and 11 months it was 67,18% in AE relation and 68,76% in BE relation; from 7 to 7 years and 11 months it was 77,61% in AE relation and 76% in BE relation; in G3, 61,99% in AE relation and 59,74% in BE relation; in G4, 49,66% in AE relation and 49,55% in BE relation. The results showed that the procedure is effective for speech recognition in adults, to children it may be effective considering the age, because the older the children are the largest the recognition rate. The procedure can become a great instrument of aid to educators in the work with individuals presenting special educational needs concerned to language problems, in the initial grades of schooling.