基于随机森林算法的服装3D人体特征识别

发布时间：2018-01-25 00:13

  本文关键词： 随机森林算法 服装D人体 次B样条 特征识别　出处：《北京服装学院学报(自然科学版)》2017年03期 　论文类型：期刊论文

【摘要】：提出了一种基于随机森林算法的3D人体特征识别算法.首先,利用bootsrap重抽样从3D人体特征样本中抽取多个样本,并对每个bootsrap样本进行建模,生成一定数量的决策树,在此基础上组合多个决策树的预测,通过投票预测特征点,把投票比例最高的点作为特征点.然后,利用3次B样条对特征点进行拟合得到3D扫描人体轮廓线,并测定人体尺寸数据.最后,将测试结果与标准测量结果进行比较,计算误差值.仿真实验表明,该方法对不同的3D扫描人体模型具有良好的识别效果.
[Abstract]:A 3D human body feature recognition algorithm based on stochastic forest algorithm is proposed. Firstly, bootsrap re-sampling is used to extract multiple samples from 3D human feature samples. Each bootsrap sample is modeled to generate a certain number of decision trees. On this basis, the prediction of multiple decision trees is combined, and the feature points are predicted by voting. The points with the highest voting ratio are taken as feature points. Then, 3D scanning human body contour is obtained by fitting the feature points with three B-spline. Finally, the human body size data are measured. The error value is calculated by comparing the test results with the standard ones. The simulation results show that the method has a good recognition effect on different 3D scanned human models.
【作者单位】： 北京服装学院信息工程学院;
【基金】：北京服装学院创新项目(120301990122)
【分类号】：TP391.41
【正文快照】： 3D人体特征识别是数字化服装研发过程中的一个关键环节,数字化服装研发的进度与质量直接决定于特征点识别的快速性和准确性.虚拟现实和数字图像处理等技术的发展,为3D人体特征的识别提供了理论依据,如何准确、快速、低成本地获取3D扫描人体的特征数据,成为当前诸多学者和科研，

本文编号：1461433