EOG-Based Calculator
A human–computer interface that lets you operate a calculator with your eyes, by classifying electrooculography (EOG) signals.
Overview
Electrooculography measures the electrical signals of eye movement. This project turns those signals into calculator input: two eye movements and a blink select a digit or an operation.
The pipeline filters raw EOG signals, compares raw-sample features against wavelet coefficients, and tunes classifiers for each class. A simulation UI highlights the hovered and selected keys in real time.
- Year
- 2025
- Area
- AI & Vision
- Role
- Signal processing & ML
Stack
- Python
- Signal Processing
- Wavelets
- Machine Learning
- Jupyter
Signal preprocessing
Band-pass filtering and artifact removal to clean blinks and muscle noise.
Feature comparison
Raw samples vs wavelet coefficients, benchmarked on classification accuracy.
Classification
Classifiers tuned with grid search for digits 0–9, the four operators and control commands.
Eye-driven UI
A calculator layout designed so each choice takes minimal eye movement.