Eye-Track-ML at Columbia Data Science Day 2025

What it is:

Poster presentation of Eye-Track-ML, an automated eye-tracking video coding pipeline, at Columbia Data Science Day

What we did:

Present "Eye-Track-ML" poster, an automated eye-tracking video coding pipeline, at Columbia Data Science Day 2025.
Demonstrate a machine learning pipeline using YOLOv11 and SAM2.1 to automate frame-by-frame coding of eye-tracking videos.

Columbia Data Science Day 2025 brochure

Yuexin Li presenting Eye-Track-ML poster

My partner, Yuexin Li, presenting (photo by me).

Read Full Project Details...

My research partner, Yuexin Li, and I presented our methods research at Columbia University’s Data Science Day on April 2nd.

We presented “Eye-Track-ML,” a machine learning pipeline developed to automate the frame-by-frame coding of eye-tracking videos. This tool increases speed of coding of gaze data, which we created to code our for our infant eye-tracking data. Our pipeline uses YOLOv11 and SAM2.1. See our poster.

We got excellent feedback and met several students interested in joining our lab, including one who has officially joined our lab from our meeting at the conference.