%0 Conference Proceedings
%T Performance of Accessible Gesture-Based Indic Keyboard
%+ Indian Institute of Information Technology, Design and Manufacturing (IIIDTM Jabalpur)
%+ Tata Consultancy Services Limited (TCSL)
%+ Indian Institute of Technology Delhi (IIT Delhi)
%+ Indian Institute of Technology Bombay (IIT Bombay)
%A Anu Bharath, Pabba
%A Jadhav, Charudatta
%A Ahire, Shashank
%A Joshi, Manjiri
%A Ahirwar, Rini
%A Joshi, Anirudha
%Z Part 3: Assistive Technology for Blind Users
%< avec comité de lecture
%( Lecture Notes in Computer Science
%B 16th IFIP Conference on Human-Computer Interaction (INTERACT)
%C Bombay, India
%Y Regina Bernhaupt
%Y Girish Dalvi
%Y Anirudha Joshi
%Y Devanuj K. Balkrishan
%Y Jacki O'Neill
%Y Marco Winckler
%I Springer International Publishing
%3 Human-Computer Interaction - INTERACT 2017
%V LNCS-10513
%N Part I
%P 205-220
%8 2017-09-25
%D 2017
%R 10.1007/978-3-319-67744-6_14
%K Accessibility
%K Indic text input
%K Visually impaired
%K Longitudinal study
%Z Computer Science [cs]Conference papers
%X Though several keyboards for Indic languages are available on Android Play store, few are accessible by the visually impaired. Particularly, none of the gesture-based keyboards are accessible. We developed an accessible prototype of the popular gesture-based, logically organised Hindi keyboard Swarachakra. In this paper, we present findings from a two-part study. In the first part, we conducted a qualitative study with 12 visually impaired users on Swarachakra. In the second part, we conducted a longitudinal, within-subject evaluation comparing Swarachakra and Google Indic keyboard. At the end of the two-week long study, 10 participants had spent an average of 6.5 h typing, including training and text input tasks. Our study establishes benchmark for text input speeds for Indic languages on virtual keyboards by visually impaired users. The mean typing speed on Swarachakra was 14.53 cpm and that on Google Indic was 12.79 cpm. The mean speeds in last session were 21.72 cpm and 18.36 cpm respectively. Regression analysis indicates that the effect of keyboard was significant. In addition, we report the user preferences, the challenges faced and qualitative findings that are relevant to future research in Indic language text input by visually impaired users.
%G English
%Z TC 13
%2 https://inria.hal.science/hal-01676172/document
%2 https://inria.hal.science/hal-01676172/file/421756_1_En_14_Chapter.pdf
%L hal-01676172
%U https://inria.hal.science/hal-01676172
%~ LORIA2
%~ IFIP-LNCS
%~ IFIP
%~ IFIP-TC13
%~ IFIP-INTERACT
%~ IFIP-LNCS-10513