Improving neuroimaging headgear placement robustness using facial-landmark guided augmented reality

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Computer Vision

Background:

• Accurate probe placement is critical for functional near-infrared spectroscopy (fNIRS) and electroencephalogram (EEG) studies.
• Operator experience and individual head shapes can compromise placement accuracy, especially in longitudinal or group studies.

Purpose of the Study:

• Develop NeuroNavigatAR (NNAR), an augmented reality (AR) and machine learning-based software.
• NNAR aims to provide real-time estimation and display of cranial landmarks for consistent headgear placement.

Main Methods:

• Utilize a facial recognition toolbox to track 3-D facial landmarks from video frames.
• Employ a precomputed transformation linking facial to cranial landmarks (nasion, preauricular points) using a large head model library.
• Render atlas-derived head landmarks onto the subject's camera stream in real-time.

Main Results:

• The open-source AR system achieves 15 frames-per-second on a laptop.
• Median 10-20 position errors were 1.52 cm (general atlas), 1.33 cm (age-matched atlas), and 0.75 cm (subject-specific surfaces).
• NNAR showed consistent prediction errors across sessions and no significant accuracy differences across age groups.

Conclusions:

• NNAR is an easy-to-use AR tool for monitoring headgear placement.
• The system is expected to significantly improve consistency and reduce setup time for fNIRS and EEG probe application.