Study of the Model for Infrared Measurement of Subcutaneous Fat
Thickness Based on SVM Algorithm
WANG Yu1, HAO Dong-mei2, AN Zheng1
1. Department of Medical Engineering,
China-Japan Friendship Hospital,
Beijing 100026, China; 2. College
of Biological Science and Biological
Engineering, Beijing University of
Technology, Beijing 100124, China
Abstract：This paper established a model for infrared measurement of subcutaneous fat thickness
based on Support Vector Machine (SVM) algorithm. 40 cases of human experiment were designed
and completed. The subcutaneous thickness of 20 body parts of each subject was measured. 6
regression models were established with nonlinear methods (including whole data, segment data,
date excepted for shoulder blade, upper limb data, lower limb data, and abdomen data). The
prediction value of each model was analyzed in comparison with the ultrasound measurement value.
The results indicated that the segment model predicted more accurately than the whole data model. The
upper limb model was most ideal, whose correlation coefficient with B ultrasound was 0.9. The results
showed the SVM model, was more suitable for accurate and rapid measurement of subcutaneous adipose
tissue thickness.
王玉1，郝冬梅2，安峥1. 基于SVM算法的红外测量皮下脂肪厚度模型研究[J]. 中国医疗设备, 2016, 31(5): 43-46.
WANG Yu1, HAO Dong-mei2, AN Zheng1. Study of the Model for Infrared Measurement of Subcutaneous Fat
Thickness Based on SVM Algorithm. China Medical Devices, 2016, 31(5): 43-46.
