Perfat: Non-contact Abdominal Obesity Assessment System

Extensive research has revealed a strong correlation between excess abdominal fat and chronic diseases, such as insulin resistance, hypertension and hypercholesterolemia. Therefore, the measurement of abdominal fat in daily life is significant. Many household body fat scales can provide abdominal (trunk) fat value for users, however, these scales require users to stand barefoot, gripping the electrodes with their arms straight, which is not friendly for those users who are unable to stand. Specialized equipment in hospitals, such as computed tomography (CT) and magnetic resonance imaging (MRI) machines, can provide accurate abdominal fat value with less restriction on user's posture, but those machines are costly, bulky and professional, making them impractical for home use. In this paper, we present Perfat, a system that aims to estimate the abdominal fat percentage accurately. The key insight of Perfat is that, as signals pass through the human body, the energy attenuation of signals at different frequencies varies, and the abdominal fat percentage can be calculated based on such attenuation variation. To evaluate the performance of Perfat, we developed a prototype using software-defined radio and conducted experiments with 50 participants. We compared Perfat with the Inbody S10 body composition analyzer and found a correlation of over 96.7%.