Research On Human Body Feature Measurement Based On Multi-parameter Fusion

Abstract

Due to the high complexity and flexibility of the human body, accurate measurement of human body size is a difficult task. The previous methods to solve these problems usually only improve on one measurement parameter and feature point markers. Such as optimization algorithm, model training, body part segmentation, point cloud processing, break the problem into multiple subtasks, and predict and calculate the size of the human body. In this article, it is recommended to fuse multiple measurement parameters, introduce SIFT algorithm and random forest method to realize the registration of feature points and the prediction of body size respectively, so that the point cloud model can be better integrated, and feature measurement from multiple directions, including: circumference, length, average value, maximum value. Finally, the feature size of the human body is calculated through comprehensive analysis, and then the accurate measurement of the three-dimensional size of the human body is realized. The proposed method is faster in calculation efficiency and higher in accuracy.