MobileGaze JS

Real-Time Gaze Estimation

RUN

This tool brings the yakhyo/gaze-estimation ONNX inference path into the EZ-MMLA Toolkit as a browser-based webcam workflow.

It keeps the same core model behavior as the upstream repository: faces are cropped, resized to 448 x 448, normalized with ImageNet statistics, passed through the gaze estimation model, and decoded into yaw and pitch angles before a gaze vector is drawn back on screen.

Within the Toolkit, it follows the same run workflow as the other JS tools: webcam capture, recording controls, downloadable CSV predictions, and an overview panel summarizing the session.

Output Description
yaw_degrees Horizontal gaze angle predicted by the ONNX model.
pitch_degrees Vertical gaze angle predicted by the ONNX model.
face_score Confidence score from the face detector used to crop the face.
face_index Per-frame face identifier for multi-person recordings.

Source Code