Prancha de comunicação alternativa por imagens empregando predição de palavras

• Ano de defesa: 2019
• Autor(a) principal: Roth, Valmir
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Curitiba; Brasil; Programa de Pós-Graduação em Engenharia Elétrica e Informática Industrial; UTFPR
• Palavras-chave em Português: Equipamentos de autoajuda para pessoas com deficiência; Pessoas com deficiência - Meios de comunicação; Processamento de imagens - Técnicas digitais; Interfaces de usuário (Sistemas de computação); Métodos iterativos (Matemática); Markov, Processos de; Algorítmos; Processamento de linguagem natural (Computação); Software - Desenvolvimento; Métodos de simulação; Engenharia biomédica; Self-help devices for people with disabilities; People with disabilities - Means of communication; Image processing - Digital techniques; User interfaces (Computer systems); Iterative methods (Mathematics); Markov processes; Algorithms; Natural language processing (Computer science); Computer software - Development; Simulation methods; Biomedical engineering; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO::SISTEMAS DE INFORMACAO; Engenharia Elétrica
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/4808

The inability to communicate with others means that people can only prevent and hinder the use of assistive technologies. In order to provide effective means of communication for these people, researchers on Alternative and Augmentative Communication (AAC) field investigate and develop new methods and technologies. Thus, in this research, we aim to develop a pictures-based communication board using word prediction to speed up information selection and minimize the user's effort to communicate effectively. The developed software has 2000 figures focused on the Brazilian culture, runs on computers running Windows 8.1 and higher versions and also smartphones and tablets running Android 7.0 and higher. With exclusive features over other free AAC applications available in Brazil, New Free Communication Board - AMPLISOFT implements a word prediction algorithm developed based on Markov's Hidden Models (Markov's Hidden Models - HMM). It is pointed by the literature as one of the most indicated word prediction techniques. A combination of the HMM word prediction algorithm and the new AAC software provides a unique experience for the user and interlocutors because it was specially designed for disabled Brazilian. For this reason, the new free communication board is flexible as to its form of use and the context in which it is intended to work. It allows different texts to be used as a basis for images prediction. And the selection of user images and expressions can be performed applying row and column scanning technique or with manual navigation. Both options are configurable in the software taking into account the corpus loaded or selected by the interlocutor. After selecting the pictures, the user also has the option of presenting the content in audio mode, thus facilitating communication. The results obtained from the tests performed on the development bench, and also the in vivo tests of the developed board, clearly evidence of the ease of use and efficiency in creating messages that can be externalized in audio form. With this application, Brazilians with severe communication difficulties can be in relation with others in an interactive and intensive way.