Two-Step Gaze Guidance
Multimodal Interfaces and Machine Learning for Multimodal Interaction(2022)
摘要
ABSTRACT One challenge of providing guidance for search tasks consists in guiding the user’s visual attention to certain objects in a potentially large search space. Previous work has tried to guide the user’s attention by providing visual, audio, or haptic cues. The state-of-the-art methods either provide hints pointing towards the approximate direction of the target location for a fast but less accurate search or require the user to perform a fine-grained search from the beginning for a precise yet less efficient search. To combine the advantage of both methods, we propose an interaction concept called Two-Step Gaze Guidance. The first-step guidance focuses on quick guidance toward the approximate direction, and the second-step guidance focuses on fine-grained guidance toward the exact location of the target. A between-subject study (N = 69) with five conditions was carried out to compare the two-step gaze guidance method with the single-step gaze guidance method. Results revealed that the proposed method outperformed the single-step gaze guidance method. More precisely, the introduction of Two-Step Gaze Guidance slightly improves the searching accuracy, and the use of spatial audio as the first-step guidance significantly helps in enhancing the searching efficiency. Our results also indicated several design suggestions for designing gaze guidance methods.
更多查看译文
关键词
guidance,two-step
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要