| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.61 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
There are several ways to input text in a computer nowadays, which the most common
is the traditional physical keyboard that has an average entry rate of 40 words per minute.
There are also, virtual keyboards that work together with a mouse, speech to text or even
touch based text input. However for people who are affected by certain motor impairments, these do not work, so their only resort is gaze text entry. Although eye interaction
has been around for many years, improvements and innovation are always possible and
necessary. Due to this relatively slow typing, in comparison to the traditional physical
keyboard, text entry could be quite boring, monotonous and uncomfortable for the user.
To this end, we needed to provide the users with a more comfortable and intuitive way of
typing that at the same time provides a good entry rate matched with a low error rate.
In this work we developed WordPop, a text entry technique to allow users to enter text
using only gaze. Our technique is based on a multi-level system of dynamic suggestions,
where users can type from one letter to a full word at a time. For the design of this
technique, we started by doing research on the literature on the topic, understanding its
problems, possible solutions and other techniques for gaze text entry.
After designing the technique, we evaluated two variants of itself to understand which
selection mechanism (Dwell-Time or Reverse Crossing) had the most promising results
and the user preference.
In the end, we finished with more user testing, this time to verify the improvement
with practice over time and how it fared against another existing technique in the same
environment. We achieved, with 5 participants, an average of 6.97 words per minute
(wpm) and a maximum of 14.60 words per minute after five sessions. Additionally, participants considered our technique more comfortable to the eyes and enjoyable to use than
the other technique. They also mentioned that it was fun to use.
Description
Tese de Mestrado, Engenharia Informática, 2022, Universidade de Lisboa, Faculdade de Ciências
Keywords
WordPop Entrada de texto Rastreamento do olhar Interação com olhar Sugestões dinâmicas Teses de mestrado - 2023