MSc Thesis Project: Stress Detection from Laptop Interaction

In this project, I built a system that detects stress passively — without wearables, sensors, or interrupting the user — just by observing how someone interacts with their laptop.

Using a combination of keystroke dynamics, mouse movement, and webcam-based gaze tracking, the model detects signs of cognitive stress with up to 89% accuracy — all from a standard laptop.

One key breakthrough was adding gaze detection using only the built-in webcam. By estimating head pose and eye movement through facial landmarks, the system infers attention direction. This led to a 30% improvement in prediction accuracy, especially for people wearing glasses.

I’m deeply grateful to my supervisor, Dr. Niall Murray, for his guidance, encouragement, and belief in the idea from the very beginning.

📘 Coming from a job background, this project also taught me the value of research thinking — asking the right questions, experimenting patiently, and digging deep into problems.

If you’re curious about how it works or where it’s headed next, I’ve written a full breakdown here:

👉 How I built a non-intrusive system to detect stress using just a laptop