Target selection by gaze pointing and manual confirmation: performance improved by locking the gaze cursor

Area of Science:

• Human-Computer Interaction
• Biomedical Engineering
• Computer Science

Background:

• Eye movement-based interactions are increasingly used in various applications.
• Combining gaze pointing with manual action is popular but faces challenges like eye jitters and premature gaze shifts.
• These issues negatively impact the accuracy and efficiency of target selection.

Purpose of the Study:

• To propose and evaluate a novel gaze-lock cursor to mitigate noise and improve target selection in eye-based HCI.
• To compare the performance of the gaze-lock cursor against a conventional gaze cursor.
• To investigate the effects of key size and gap on typing performance with the proposed cursor.

Main Methods:

• A gaze-lock cursor was developed to keep the cursor at the target for manual confirmation.
• A typing task was conducted comparing the gaze-lock cursor with a conventional gaze cursor.
• Key size and key gap were varied to assess their impact on performance metrics.

Main Results:

• Typing performance improved significantly with larger key sizes and gaps.
• The gaze-lock cursor demonstrated a significant increase in speed and a decrease in errors compared to the conventional gaze cursor.
• The proposed gaze-lock cursor is effective in improving accuracy and efficiency in eye-based target selection.

Conclusions:

• The gaze-lock cursor is a promising advancement for eye-based human-computer interactions, particularly for tasks requiring frequent target selection.
• Increasing target size and gap further enhances performance in gaze-based selection tasks.
• The findings suggest practical improvements for designing more robust and user-friendly eye-tracking interfaces.