Accessibility

Demo Video

Overview

This Node-RED accessibility flow takes written documents and converts them to the spoken word for the visually impaired.

The flow uses the Camera and Play Audio nodes with the Watson Visual Recognition and Text-to-Speech services to enhance system accessibility.

Node-RED will:

• Capture an image from the device camera
• Recognise any text in the image
• Extract the interpreted text
• Convert the text to spoken word
• Play the audio output from the device's speakers

Application flow

Link to Flow JSON

Flow description

• The camera node will capture a still image from the device webcam when you click the button.
• The visual recognition service takes the buffer sent from the camera node and parses it to determine if there are any words in the image.
• Next the flow calls move msg.result to retrieve the string from the visual recognition service and set that to the msg.payload object required by the rest of the flow.
• Watson Text-to-Speech service then takes the string interpreted from the image and converts it to the spoken word
• The flow then sets the msg.payload object required for the play audio node.
• Finally the play audio node broadcasts the spoken word from the device's speakers

Gotchas

• The camera node has a 2000ms delay to prevent slow camera driver startup causing an issue in some browsers
• The Visual Recognition service is currently in dark beta, as a result there may be breaking changes without warning. We'll do our best to keep up with these changes - but if you find a bug please let us know by submitting an issue against this GitHub repo.