The application aims to facilitate communication for people with speech and movement disabilities, especially those who do not speak and require assistance in communication.
And Look to Speak allows users to select pre-defined phrases on the phone screen with their own eyes, in order to express themselves in an easier way.
Look to Speak uses machine learning and eye-squinting technology to choose from a list of phone phrases to speak out loud.
How said: Given that mobile devices have become very popular, and in the presence of Technologies, like machine learning, within them, have pondered ways in which phones could work with assistive technologies.
He believes how Look to Speak could be used as an extension of the existing eye-stare technology.
He added: We don’t replace these types of communication aids because they have a lot of functionality, and we use Look to Speak for important SMS.
The application aims to work via the mobile device with the support of advanced assistive technologies, and after placing the phone slightly below eye level, the user looks to the left or right to choose from the list of phrases, which the device speaks aloud.
Phrases can be customized, and according to Google, all data in Look to Speak is private and stays within the phone.
One drawback is that the application’s settings menu cannot be accessed. Phrases can be modified and squinting settings can be adjusted, through eye-stare technology, and requires clicking manually across the phone screen.
The app uses the front camera of the device to track the user’s point of view, the gaze sensitivity can be adjusted, and the app falls asleep while looking up.
How says: The app has reached people who can benefit from it and have found it useful in environments where other access devices may not be usable, such as: during transport, in the shower, in urgent situations, or outdoors.
The app comes as part of the Start with One project under the Experiments with Google umbrella, and other projects include Teachable Machine, a web tool that allows users to create machine learning models without programming.