Fully Automated Pipeline for Body Composition Estimation from 3D Optical Scans using Principal Component Analysis: A...

Area of Science:

• Biomedical Engineering
• Anthropometry
• Health Informatics

Background:

• 3D optical scan (3DO) technology is increasingly used for body composition and health risk prediction.
• Current methods often rely on manual point placement, which is time-consuming and expensive.
• The Shape Up studies aim for a device-agnostic solution using principal component analysis (PCA).

Purpose of the Study:

• To develop and validate an automated landmark detection method for 3DO scans.
• To enable a fully automated, device-agnostic PCA-based approach for body composition estimation.
• To assess if automated landmarks maintain the quality of body composition estimates compared to manual methods.

Main Methods:

• Implementation of a novel automated landmark detection algorithm for 3DO scans.
• Application of Principal Component Analysis (PCA) on landmark data for body composition prediction.
• Comparison of body composition estimates derived from automated versus manual landmark placement.

Main Results:

• Proof-of-concept for a fully automated PCA-based body composition estimation pipeline.
• Automated landmark detection successfully replaces manual point placement without compromising estimate quality.
• The proposed method facilitates a more practical and scalable solution for real-world applications.

Conclusions:

• Automated landmark detection is a viable and effective method for 3DO-based body composition analysis.
• This innovation enhances the practicality of device-agnostic PCA solutions for health risk assessment.
• The developed pipeline supports efficient and accessible body composition estimation in diverse settings.