Machine Learning Intern, Eye Tracking Data Labeling Project
Engineered a real-time eye tracking model using MediaPipe and various machine learning algorithms to minimize RMSE. Collected and annotated 478 facial landmarks as labeled data points for precise model training. Calibrated and validated the tracking interface using labeled keypoint data to enhance accuracy and user experience. • Labeled facial landmark keypoints for eye tracking in image frames. • Performed feature selection to optimize model training with annotated data. • Developed a data collection and labeling workflow for calibration. • Used internal/proprietary tooling in a Python/Tkinter-based interface.