Improving Gaze-based Selection using Variable Dwell Time.
Year: 2017
Type: preprint
Abstract: In order to avoid the Midas Touch problem, gaze-based interfaces for selection often introduce a dwell time: a fixed amount of time the user must fixate upon an object before it is selected. Past interfaces have used a uniform dwell time across all objects. Here, we propose an algorithm for adjusting the dwell times of different objects based on the inferred probability that the user wishes to select them. In particular, we introduce a probabilistic model of gaze behavior and train it on natural gaze trajectories collected while users surf the web through a gaze-based web browser. Using this model, we can infer the probability that each hyperlink on the web page is the one the user intends to select. We assign shorter dwell times to more likely hyperlinks and longer dwell times to less likely hyperlinks. The resulting variable dwell time gaze-based browser enables subjects to surf the web faster and/or more reliably. We use a two-stage hidden Markov model based algorithm to model the natural eye gaze trajectories. We have evaluated this method objectively both in simulation and experimentally, and subjectively through questionnaires. Our results demonstrate that the proposed algorithm achieves a better tradeoff between accuracy and speed. This work demonstrates that natural eye movements can be used to infer user intent, and that this information can be used to improve gaze-based selection without increased cognitive load on users.
Source: arXiv (Cornell University)
Authors: Zhaokang Chen, Bertram E. Shi
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2609705732