PAROL+ is an innovative application for tablets that allows people who have lost speech (tracheotomy, Amyotrophic Lateral Sclerosis) to communicate simply. The application aims at replacing the lost voice by enabling the user to compose a message using a smart text-to-speech system.
PAROL+ allows to compose messages messages that can be heard orally by a voice synthesis, or directly shared by email, text or any other messaging tools.
PAROL+ first goal is to facilitate message composition.
To help keystroke saving, we developed Mind2speech, a word prediction engine. It calculates the words needed by the user at every step of message composition.
The calculus relies on statistical observations over textual data. The prediction engine tunes itself to adapt to the user’s lexicon and sentences over time. It improves the predictive precision and the speed of composition of a message (by predicting/proposing the most relevant/probable words for the rest of the message).
PAROL+ keeps a record of the messages. The user can save emergency messages or bookmark frequent ones that can be quickly re-used.
The predictions are presented in list form. If the word is present in the list, only one gesture is needed to write it. When the user can’t find the word on the list, access to the keyboard is possible. A list of dynamic completions is then proposed according to the characters entered.
The predictions can be sorted and organized in the distinctive color zones for the different criteria: most probable words first, alphabetical order, word categories (verbs, names, etc.).