Abstract
Browser-based webcam gaze trackers are increasingly used for crowd-scale data collection and in clinical settings where lab eye
trackers are impractical, but the reported latency numbers may not represent real world functionality. The common practice of timestamping each gaze sample when it is emitted, rather than when its source frame was captured, makes the measured inference latency read about $0\,$ms no matter how slow the engine really is. We show how to measure it correctly, recovering a true per-frame capture clock from the browser’s \texttt{re-quest-Video-Frame-Call-back} (rVFC) API: exact pairing through a per-frame queue for engines that expose their inference pipeline, and a verifiable lower bound for engines that do not, such as WebGazer. We release an open TypeScript implementation and benchmark harness, demonstrated on two interchangeable engines: WebGazer and a new FaceMesh+KRR pipeline. On commodity hardware (N=1), correct timestamping raises the reported inference latency from about $0\,$ms to a $22$–$34\,$ms median ($27$–$52\,$ms $p_{95}$), a $20$–$50\,$ms gap that is enough to change whether a $50\,$ms interactive-latency budget is met. The harness also separates two notions of precision that aggregate error conflates: the spatial spread of a fixation cluster and its temporal jitter. Cross-engine accuracy differences sit within the $\sim 7^\circ$ single-session noise floor we measure by ablation, so we treat them as observations, not a ranking. Finally, we feed the webcam gaze into a published clinical weak-supervision pipeline (GazeMedSeg on Kvasir-SEG): with only the gaze source changed, expert eye-tracker gaze trains a usable polyp segmenter (Dice $0.68$) and our webcam gaze does not (Dice $\approx 0$). At this accuracy, fine lesion labelling is out of reach. A multi-user replication is in preparation.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/DEMI_025.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=VNRuJI79iq
BibTex
@InProceedings{CheChi_Measuring_MICCAISAT2026,
author = { Chen, Chi-Sheng AND Brat, Gabriel A.},
title = { { Measuring Browser Webcam Gaze Honestly: A Capture-Clock Methodology and Open Reference Implementation } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17276},
month = {pending},
page = {pending}
}