Improving FlexType: Ambiguous Text Input for Users with Visual Impairments

Home > Publications > Improving FlexType: Ambiguous Text Input for Users with Visual Impairments

Improving FlexType: Ambiguous Text Input for Users with Visual Impairments

Dylan Gaines, Keith Vertanen

PETRA '24: Proceedings of the 16th International Conference on PErvasive Technologies Related to Assistive Environments, 2024 (to appear).

We present an improved version of the FlexType interface developed in our previous work. FlexType enables nonvisual text input on mobile touchscreen devices by allowing users to select from a small number of character groups with gestures instead of targeting letters at specific screen locations. Based on an interview with users who are blind or low vision, we added a letter-entry mode to enable any word to be typed, even if it is not recognized by the decoder. We conducted a longitudinal study with users who are legally blind to compare FlexType to users' typical text input methods. While we found FlexType was significantly slower than Braille Screen Input, there was no significant difference in entry or error rate between FlexType and an onscreen keyboard with VoiceOver enabled. This was despite the participants having much more experience with the VoiceOver keyboard compared to FlexType. Overall, participants averaged 7.7 words per minute with a 7.0% character error rate when writing with FlexType.

Reference: