Text Entry

Soft keyboards

DUCK Keyboard DUCK aims reducing tactile exploration and speeding up text entry for visually impaired users. It relies on the selection of the first letter of a word and rapid and inaccurate typing of the remaining letters. It then proposes a list of words having the same first letter and a similar total distance between letters. Main publication:

Semantic Keyboard Semantic Keyboard is a soft keyboard augmented by semantic pointing. The cursor crosses faster over keys containing low-probability letters (considering the prefix already entered). This optimization reduces the movement of the pointer by 60\%, and increases the text entry speed by 13.5\%. Accuracy is equivalent to a regular soft keyboard. Main publication:

Fisheye Keyboard Our goal is to find a compromise between readability and usability on a whole character layout for a smartphone. The proposed interactions allow to display all keys on a small screen while making pointing easier for the user by expanding any given key as a function of its distance from the pointer. Main publication:

Floodkey Floodkey is a sof keyboard in which some keys are increased in size while needless ones are decreased. Each time a key is pressed, according to the probabilities of the key to be pressed, the sites are located by the Fruchterman-Reingold algorithm that maintains the topology of the keyboard. Then the keyboard appears with no overlapping as a weighted Voronoi diagram.

SpreadKey The keyboard is based on a QWERTY layout and dynamically recycles the needless characters. The needless characters are determines by a predictive system in function of the previous input characters. The recycling aims at providing several times the characters with a high probability to be typed. Main publication:

KeyGlass System KeyGlass System is based on the dynamic addition of semi-transparent keys that are given according to previously entered characters and a predictive system. The prediction system that we use to optimize our system is based on the joint use of a lexicographic tree and a system using bigrams.

GAG Keyboard We propose a genetic algorithm formal framework to optimize character location on a soft keyboard. This method is described regardless of the language and layout used and can then easily be adapted to any language and layout. Main publication:

Prediction systems

Word List Optimization for Visually Impaired UsErs on Smartphones LOVIE is a List Optimized for Visually Impaired users. We conduct three studies in order to find which interaction would be the most efficient for blind navigation and selection: 1) on the navigation in the word lists, 2) on the audio feedback and 3) on the validation step. Our goal is to reduce the time taken to select words in a list, without knowing the items beforehand (because it is in a dynamic context).

Centralist Centralist is a predictive text entry system with a new character layout around a word list. Our goal is to optimize the use of software keyboards associated with the prediction list. Our new design allows the user to keep focus on the keyboard and the list, as well as reduce cursor movement.

Wordtree WordTree is a list of words in which the user can select a substring of the word. The user clicks on any letter, and the substring to this letter is inserted in the text.