Using a marker-less method for estimating L5/S1 moments during symmetrical lifting.

Area of Science:

• Biomechanics
• Occupational Health
• Computer Vision

Background:

• Assessing L5/S1 joint moments is crucial for identifying non-ergonomic lifting tasks.
• Current laboratory-based methods are impractical for on-site biomechanical analysis.
• A need exists for accessible tools to evaluate lower back loads during occupational lifting.

Purpose of the Study:

• To analyze the validity of a computer vision-based method for estimating 3D L5/S1 joint moments.
• To assess the reliability of this novel method for occupational biomechanical analysis.
• To provide a practical alternative to marker-based motion tracking systems.

Main Methods:

• A computer vision-based method was developed to estimate 3D L5/S1 joint moments during symmetrical lifting.
• External validation compared the computer vision method against a lab-based reference method.
• Internal validation compared top-down and bottom-up biomechanical models.

Main Results:

• No significant differences were found in peak and mean L5/S1 joint moments between the computer vision method and the lab-based reference.
• The computer vision method demonstrated excellent reliability, with intra-class correlation coefficients exceeding 0.91.
• The proposed method proved to be a reliable tool for assessing lower back loads.

Conclusions:

• The computer vision-based method is a valid and reliable tool for estimating 3D L5/S1 joint moments during lifting.
• This method offers a practical alternative for on-site biomechanical analysis, especially when marker-based systems are not feasible.
• The findings support the use of this technology for improving occupational safety and reducing the risk of lower back injuries.