Dry Electrode-Based Body Fat Estimation System with Anthropometric Data for Use in a Wearable Device

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Body Composition Analysis

Background:

• Bioelectrical impedance analysis (BIA) is a common method for estimating percent body fat (PBF).
• Traditional BIA methods often require multiple electrodes and lengthy measurement times.
• Existing wearable devices face challenges with accuracy and measurement consistency due to parasitic resistance and posture variations.

Purpose of the Study:

• To develop a faster, more accurate wearable device for PBF prediction using BIA.
• To utilize upper-body impedance to predict whole-body impedance, simplifying the measurement process.
• To minimize measurement variations through a novel sensing system and reduce measurement time.

Main Methods:

• A small, dry electrode-based wearable device was designed, focusing on wrist and index finger contact.
• A parameter calibration network was employed to reduce measurement time to five seconds.
• A deep neural network algorithm integrated upper-body impedance with lower-body anthropometric data for PBF prediction.

Main Results:

• The proposed system demonstrated improved performance compared to two commercial BIA systems.
• A significant improvement in the correlation coefficient (r²) of approximately 9% was observed.
• The standard error of estimate (SEE) was reduced by 28%, indicating enhanced prediction accuracy.

Conclusions:

• The novel wearable BIA device offers a more efficient and accurate method for PBF assessment.
• The deep neural network approach effectively predicts PBF using limited impedance data and auxiliary features.
• This technology has the potential to enhance personal health monitoring and athletic performance tracking.