The People's Gaze: Co-Designing and Refining Gaze Gestures with General Users and Gaze Interaction Experts

A research team led by Yaxiong Lei has introduced a novel, user-centered methodology for designing intuitive gaze gestures, detailed in their paper 'The People's Gaze,' accepted for ACM CHI 2026. The work directly addresses a critical gap in human-computer interaction: as eye-tracking becomes standard in mobile devices, existing gesture sets are often designed by experts and fail to align with how people naturally move their eyes. To solve this, the researchers developed a two-phase co-design process.

In the first phase, they conducted four workshops with 20 non-expert participants, who generated 102 initial gesture concepts after a brief introduction. The study found that users intuitively anchored their ideas in familiar metaphors and developed a compositional 'grammar' for interaction: an activation step (like a dwell) followed by an action (a gaze path or blink). This structure is key to ensuring intentionality and mitigating the classic 'Midas Touch' problem, where a system misinterprets every glance as a command.

In the second phase, four gaze interaction experts reviewed and refined the user-generated concepts. They prioritized gestures that were ergonomically sound, aligned with natural eye movements (saccades), and could be reliably distinguished from one another by tracking systems. This expert validation honed the broad concepts into a practical, finalized set of 32 gestures.

The resulting framework offers more than just a new gesture library; it provides actionable design principles for the industry. By grounding the design in real user intuition and then vetting it for technical feasibility, the research lays a foundational blueprint for developing the next generation of seamless, hands-free interfaces for AR/VR headsets, smartphones, and assistive technologies.