AI for Eyetracking

Full event details coming soon — please save the date!

Title: AI for Eyetracking

Date & Time: Thursday, December 4, 2025 | 6:30 – 8:30 PM

Duration: 2 Hours

Description:
🔹 1. Eye & Gaze Tracking

Old approach (non-AI): Relied on infrared + geometric models of the eye to calculate gaze. Very precise but hardware-heavy.

Now (AI-powered): Uses deep learning gaze estimation models trained on eye images.

Example: WebGazer.js uses ML to track gaze with just a webcam.

Tobii and Pupil Labs integrate AI for calibration-free and adaptive tracking.

🔹 2. Facial Tracking

Old approach: Landmark detection (geometry-based) → track nose, eyes, mouth with simple math.

Now: AI/ML models detect facial landmarks, micro-expressions, and emotions with much higher accuracy.

MediaPipe Face Mesh (by Google) uses neural nets.

Affectiva uses deep learning to recognize emotional states.

🔹 3. Object Tracking

Old approach: Color filters, optical flow, bounding box matching.

Now: AI models like YOLO, Detectron, DeepSORT track objects robustly in real-world conditions (cars, people, animals).

🔹 4. Body & Pose Tracking

Old approach: Motion sensors or reflective markers.

Now: AI pose estimation (OpenPose, MediaPipe, DeepMotion) learns body joint positions from millions of images/videos → no sensors needed.

🔹 5. Behavior Tracking

Now: Purely AI-driven — models analyze eye, face, and body data together to infer engagement, attention, fatigue, or emotion.

Platforms like iMotions, Affectiva, Realeyes rely on machine learning pipelines.

What to Bring:
To participate, bring your laptop with Figma and GPT set up and ready to go!

Location:
NYIT, 16 W 61st St, NY, NY 10023 (11th Floor Auditorium).
Light drinks will be provided

Hosted by: NYIT + Kevin & the NYC UX & Digital Product Designers Team