Enhancing gaze interaction performance: Design and optimization of perspective-driven fisheye view

Abstract

Gaze interaction has been receiving increasing attention. However, it faces the critical challenge of low spatial accuracy. To enhance the performance of gaze interaction in high-information-density interfaces, this study explores the effects of fisheye views and their key parameters (max scale and effective range) on gaze interaction. To maintain consistency in motor space, the study proposes a perspective-driven fisheye view algorithm. Based on this algorithm, a gaze interaction experiment involving single-character visual search was conducted. The results indicate that fisheye views significantly improve both temporal and spatial performance in gaze interaction without altering the size of the motor space. Furthermore, the max scale and effective range of the fisheye view have varying impacts on different performance metrics. Overall, a max scale of 175 % and an effective range of 3 times the element center distance are identified as optimal parameter settings. These findings provides important theoretical and practical guidance for the application of fisheye view, and underscore the critical role of visual attention guidance in gaze interaction.