Estimating Lumbar Spine Compression Using Markerless Motion Capture

Compressive forces at the L5-S1 intervertebral joint are a contributing factor to low back pain, a leading cause of work-related musculoskeletal disorders. Estimating these forces in field settings and comparing them to safety limits can support ergonomic risk assessments. While musculoskeletal modeling platforms such as OpenSim provide accurate estimations of spinal loading, their reliance on expert operation and high computational demands limits their practicality in field settings. This study presents a biomechanical model that estimates L5-S1 compression force using joint angles derived from a markerless motion capture system, along with the user’s body weight, sex, and the lifted load. A proof-of-concept evaluation was conducted with one participant performing symmetric lifting tasks at 5, 10, and 15 kg. The proposed method’s outputs were compared to OpenSim estimates, showing close agreement in peak compression force for 10 and 15 kg conditions, with normalized peak estimation errors of 4 ± 2% and 2 ± 1% during lifting and 11 ± 5% and 5 ± 4% during lowering, respectively. Larger errors were observed for the 5 kg condition and during mid-movement for all load conditions, likely due to unmodeled muscle co-contraction and static assumptions. Overall, the method shows promise for accessible in-field ergonomic assessment of peak spinal loads during lifting and provides a foundation for evaluating interventions such as occupational exoskeletons in real-world settings.