Scholarlink

The fashion, healthcare, and fitness industries are becoming increasingly reliant on accurate anthropometric measurements of the human body. This paper aims to automate the extraction of precise 3D body measurements for different ana-tomical regions of the upper body. The proposed method involves converting 2D images into 3D meshes, utilizing computer vision techniques and 3D modelling software to automatically extract key anthropometric dimensions. By applying mathematical formulae, we calculate various circumferences and distances, including chest, waist, shoulders, and elbow measurements. Our unique dataset of 15 individuals enabled us to identify critical physical landmarks relative to the body’s height. The mean absolute errors (MAE) for the chest, waist, shoulders, left elbow, and right elbow measurements are 6.17 inches, 6.05 inches, 2.72 inches, 0.951 inches, and 0.9803 inches, respectively. These results demonstrate the accuracy and effectiveness of the proposed methodology for predicting significant body dimensions in 3D space.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automation of Human Body Measurements from 2D Images

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Chaitanya Srikanth,
  • Sumanthra Rao Yerabelly,
  • Pradham Mummaleti,
  • Ankita Jain

摘要

The fashion, healthcare, and fitness industries are becoming increasingly reliant on accurate anthropometric measurements of the human body. This paper aims to automate the extraction of precise 3D body measurements for different ana-tomical regions of the upper body. The proposed method involves converting 2D images into 3D meshes, utilizing computer vision techniques and 3D modelling software to automatically extract key anthropometric dimensions. By applying mathematical formulae, we calculate various circumferences and distances, including chest, waist, shoulders, and elbow measurements. Our unique dataset of 15 individuals enabled us to identify critical physical landmarks relative to the body’s height. The mean absolute errors (MAE) for the chest, waist, shoulders, left elbow, and right elbow measurements are 6.17 inches, 6.05 inches, 2.72 inches, 0.951 inches, and 0.9803 inches, respectively. These results demonstrate the accuracy and effectiveness of the proposed methodology for predicting significant body dimensions in 3D space.