Finger Based Techniques for Nonvisual Touchscreen Text Entry

TL;DR

This paper introduces finger-based techniques for non-visual touchscreen text entry for blind users, demonstrating improved speed and accuracy with specific models and identifying accessible regions for widget placement.

Contribution

It proposes novel finger-based interaction models for blind users and provides a usability framework for accessible widget placement on touchscreens.

Findings

Single Digit FDI is faster and more accurate than Double Digit FDI and voice-over.

Finger-Text Input (FTI) shows promising results for text entry.

11 accessible regions identified for widget placement to aid blind users.

Abstract

This research proposes Finger Based Technique (FBT) for non-visual touch screen device interaction designed for blind users. Based on the proposed technique, the blind user can access virtual keys based on finger holding positions. Three different models have been proposed. They are Single Digit Finger-Digit Input (FDI), Double Digit FDI for digital text entry, and Finger-Text Input (FTI) for normal text entry. All the proposed models were implemented with voice feedback while enabling touch as the input gesture. The models were evaluated with 7 blind participants with Samsung Galaxy S2 apparatus. The results show that Single Digit FDI is substantially faster and more accurate than Double Digit FDI and iPhone voice-over. FTI also looks promising for text entry. Our study also reveals 11 accessible regions to place widgets for quick access by blind users in flat touch screen based smartphones. Identification of these accessible regions will promote dynamic interactions for blind users and serve as a usability design framework for touch screen applications.