FAAVSP - The 1st Joint Conference on Facial Analysis, Animation, and
Auditory-Visual Speech Processing

Vienna, Austria
September 11-13, 2015

Comparison of Visual Speech Perception of Sampled-Based Talking Heads: Adults and Children With and Without Developmental Dyslexia

Paula D. Paro Costa (1), Daniella Batista (1), Mayara Toffoli (1), Keila A. Baraldi Knobel (1), Cíntia Alves Salgado (2), José Mario De Martino (1)

(1) University of Campinas; (2) Federal University of Rio Grande do Norte, Brazil

Among other applications, videorealistic talking heads are envisioned as a programmable tool to train skills which involve the observation of human face. This work presents partial results of a study conducted with adults and children with and without developmental dyslexia to evaluate the level of speech intelligibility provided by a sample-based talking head model in comparison with unimodal auditory and real video stimuli. The results obtained so far indicate that dyslexic children, when compared to control group, are less capable of dealing with the imperfections of a synthetic facial animation visual speech model. The present study is motivated by the hypothesis that a training program using facial animation stimuli could improve the phoneme awareness of dyslexic children, a skill that is considered significantly related to success in the early stages of reading and spelling. Index Terms: dyslexia, facial animation, speech intelligibility, assistive technology

Full Paper

Bibliographic reference.  Paro Costa, Paula D. / Batista, Daniella / Toffoli, Mayara / Baraldi Knobel, Keila A. / Alves Salgado, Cíntia / Martino, José Mario De (2015): "Comparison of visual speech perception of sampled-based talking heads: adults and children with and without developmental dyslexia", In FAAVSP-2015, 12-16.