Skip to Main content Skip to Navigation

Techniques d'assistance à la saisie de texte sur périphériques mobiles dans le cas de la déficience visuelle

Abstract : Over the past 10 years, mobile phones have evolved considerably : the advent of touch screens and the disappearance of physical keyboards have changed the way we interact with these devices in our daily lives. However, text input is still an important task though activities such as taking notes, sending text messages or communicating on social networks. However, while using touch screen has great advantages in terms of dynamical interfaces and customization, such devices are not necessarily accessible to all. Indeed, for 39 million people worldwide affected by blindness, there are many difficulties related to the devices that will make such interactions difficult due to the lack of tactile reference: interactions with the device are possible but they are often laborious and repetitive, which then implies a lot of cognitive load, accuracy, memory and fatigue related problems. In this PhD thesis, we focused on the accessibility of text input interactions in the context of visual impairment. First, we studied the currently existing solutions designed for impaired users. The main issue of this research was to improve text input so that users have a better typing experience. As such, we designed a deductive solution called DUCK. This solution allows visually impaired to quickly enter text without worrying about how accurate their input is. A basic language-based modelling system allows at the each end of a word what the user wanted to type. This device was then tested with a sample of visually impaired people to assess the effectiveness of our solution. Further work was subsequently focused on two main optimizations. The first work focused on word lists. We studied and compared different interactions to allow the user to navigate and choose words effectively and easily when faced with a list of words proposed by a predictive or deductive system. The second optimization focuses on entering commonly used words. We also conducted a comparative study of different interactions models to type a short word efficiently without using the deduction system, which would be too time consuming for such words. Finally, we finish this PhD project by a longitudinal study that shows the DUCK keyboard with the integration of these optimizations. This new system has been used by visually impaired over a period of two weeks to study the effectiveness of the keyboard over the long term.
Document type :
Complete list of metadatas

Cited literature [58 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Monday, March 26, 2018 - 4:14:50 PM
Last modification on : Tuesday, October 20, 2020 - 11:30:33 AM
Long-term archiving on: : Thursday, September 13, 2018 - 9:25:14 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01743690, version 1



Philippe Roussille. Techniques d'assistance à la saisie de texte sur périphériques mobiles dans le cas de la déficience visuelle. Informatique mobile. Université Paul Sabatier - Toulouse III, 2017. Français. ⟨NNT : 2017TOU30020⟩. ⟨tel-01743690⟩



Record views


Files downloads