CoGaze · Takuma Masuda
Takuma Masuda
UIST 2026 · Poster First author NUS · AHLab

CoGaze: Webcam-Based Gaze Sharing for Asymmetric Remote Collaboration

On a video call you cannot say “that one” and just look at it. CoGaze puts one person's gaze on the other person's screen, estimated from the webcam they already have.

UIST video, 30s.

Venue
UIST 2026, Poster Session (Adjunct)
Role
First author. Built the system, ran the study
Where · when
Augmented Human Lab, NUS · Aug – Nov 2025
Supervision
Prof. Suranga Nanayakkara, Dr. Prasanth Sasikumar

The problem

In asymmetric remote collaboration, a remote helper guides an on-site worker through a video link. They never share a viewpoint, so neither can point with their eyes the way two people at the same table do without thinking.

A mouse cursor is the usual substitute, and it is a poor one. The helper's hands are often already busy, and a cursor only moves once you have decided to point. Your eyes were there a second earlier.

Why a webcam

A lab eye-tracker is precise and still rare. A webcam is on every laptop and misses by a few degrees. Eye tracking will spread, but unevenly, so collaborators will keep arriving with devices that do not match. CoGaze estimates gaze from a consumer webcam on one side, and asks which kinds of shared attention survive that mismatch.

What we compared

The study compared three ways of drawing someone else's gaze on screen: a ray, a frustum and a circle.

Three ways of presenting eye gaze: ray, frustum and circle

What is next

CoGaze was a small version of a larger question: can one person hand their attention to another without replacing what the other person sees? Shared gaze was already known to work in MR when both people wear the same headset. CoGaze looked at a pairing that does not match. The open question is how attention should be shown when the two sides sense and see differently, whatever the hardware becomes.

Primary materials

Paper
ACM Digital Library ↗
doi:10.1145/3830397.3841904
Video
UIST video on YouTube ↗
30 seconds
Paper PDF
Author copy ↓
UIST Adjunct '26
Cite

Takuma Masuda, Prasanth Sasikumar, Koichi Hirota, Sho Sakurai, Takuya Nojima, Suranga Nanayakkara. 2026. CoGaze: Webcam-Based Gaze Sharing for Asymmetric Remote Collaboration. In Adjunct Proceedings of UIST '26. ACM. doi:10.1145/3830397.3841904

{{ bib }}