09-03-2017, 01:07 PM
A new form of augmentative and alternative communication device (AAC) for people with severe speech disabilities-voice input-output communication assistance (VIVOCA). VIVOCA recognizes the messy discourse of the user and constructs messages, which become synthetic speech. The development of the system was carried out using user-centered design and development methods that identified and refined the key requirements for the device. A new methodology was applied for the construction of small vocabulary, speaker-dependent automatic speech recognizers with reduced amounts of training data. Experiments showed that this method succeeds in generating a good recognition performance (mean 96% accuracy) in highly disordered speech, even when recognition perplexity increases. The selected message construction technique interchanged several factors including the message build speed and the range of available message outputs. VIVOCA was evaluated in a field trial by individuals with moderate to severe dysarthria and confirmed that they can make use of the device to produce intelligible speech output from disordered speech input. The trial highlighted some issues that limit the performance and usability of the device when applied in real-life situations, with a recognition accuracy of 67% under these circumstances. These limitations will be addressed in future work.