Objective Dynamic Assessment of Facial Movement Asymmetry in Children Using a Marker-Based Video Method.

Area of Science:

• Biomedical Engineering
• Neurology
• Computer Vision

Background:

• Facial movement symmetry is crucial for assessing neuromuscular function, with asymmetries indicating potential neurological disorders, trauma, or post-surgical complications.
• Quantitative symmetry assessment is vital for accurate diagnosis, effective therapy monitoring, and precise surgical planning in clinical practice.

Purpose of the Study:

• To develop and validate a marker-based approach for stable tracking of facial movements.
• To investigate the efficacy of dynamic facial movement descriptors in differentiating symmetric from asymmetric exercise execution.

Main Methods:

• A low-cost video acquisition setup was used with 75 patients performing eyebrow raising and smiling exercises.
• Seventeen ArUco markers were placed on facial landmarks, and a processing pipeline involving marker/face detection and symmetry axis estimation was employed.
• Dynamic descriptors, VertDist (vertical displacement) and Ratio (relative position), along with their derivatives, were extracted and statistically analyzed.

Main Results:

• Statistically significant differences between symmetric and asymmetric movements were identified, particularly in the first derivatives of VertDist and Ratio.
• Large effect sizes were observed for VertDist in eyebrow raising and for Ratio in smiling.
• Moderate effect sizes were noted for Ratio in eyebrow raising and VertDist in smiling.

Conclusions:

• The proposed marker-based method provides a reliable and cost-effective solution for quantifying facial movement asymmetry.
• Dynamic descriptors derived from VertDist and Ratio show significant potential in distinguishing symmetric from asymmetric facial movements, supporting clinical applications.