Single-Camera Knee Adduction Moment Estimation for Individuals With Knee Osteoarthritis via a Novel Spatio-Temporal Graph Transformer Network.

The proposed STGTN model achieved KAM estimation accuracy within clinically meaningful error thresholds using a single-camera setup.

• KAM root mean square error was 0.48% BW·BH
• Peak KAM mean absolute error was 0.43% BW·BH
• Both metrics fell within the clinically meaningful error threshold range of 0.5–2.1% BW·BH
• The model was tested on 14 individuals with medial compartment knee OA

The STGTN model demonstrated sensitivity to gait modifications by identifying significant reductions in peak KAM across multiple conditions.

• Significant reductions in peak KAM were identified during slow walking, toe-in gait, wide step width, and increased trunk sway conditions
• All identified reductions were statistically significant (p < 0.05)
• Gait modifications tested included variations in walking speed, foot progression angle, step width, trunk sway, and dual-task walking

Fourteen individuals with medial compartment knee osteoarthritis performed multiple gait modifications as the study population.

• Participants had medial compartment knee OA
• Five categories of gait modifications were performed: walking speed, foot progression angle, step width, trunk sway, and dual-task walking
• This represents a diverse set of gait conditions for model validation

Traditional KAM measurement systems are costly and restricted to laboratories, motivating the development of a single-camera alternative.

• Traditional motion capture and force plate systems are costly and restricted to laboratories
• Multi-wearable sensor and multi-camera setups have been explored but remain complex
• Prior multi-sensor and multi-camera approaches lack validation across diverse gait modifications in individuals with knee OA
• A single-camera setup was proposed as a feasible approach for real-world gait assessment

The single-camera-based KAM estimation approach has potential applications in clinical and rehabilitation settings for monitoring knee joint loading and exploring feedback-driven KAM reduction.

• The approach is described as potentially applicable for monitoring knee joint loading
• The authors highlight potential for feedback-driven KAM reduction applications
• The system targets real-world gait assessment outside of traditional laboratory settings
• The model's sensitivity to clinically relevant gait modifications supports its translational potential