3D Human Body Model Reconstruction Algorithm Based on Multi-View Synchronized Video Sequences

Yidong Wu

2024

Abstract

As an important topic in the field of computer vision, 3D human reconstruction has a wide range of applications in the fields of film and television entertainment, sports and medicine. Traditional 3D human reconstruction methods often require professional equipment and clothing for technical support, and the process is very cumbersome and has great limitations. In recent years, with the development of deep learning, the method of human reconstruction using deep learning has achieved great success. Based on this background, this paper introduces a 3D human reconstruction algorithm based on multi-view synchronized video sequences, which can improve the shortcomings of traditional methods. Specifically, this paper reprojects the key points on the 3D human model back to the 2D plane under multiple perspectives, and uses the key points obtained by 2D human posture detection to optimize the reprojected key points, and finally obtains the body shape and posture parameters of the 3D human model. After comparative experiments, the method of this paper has achieved good accuracy and efficiency.

Paper Citation

in Harvard Style

Wu Y. (2024). 3D Human Body Model Reconstruction Algorithm Based on Multi-View Synchronized Video Sequences. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 214-220. DOI: 10.5220/0013512800004619

in Bibtex Style

@conference{daml24,
author={Yidong Wu},
title={3D Human Body Model Reconstruction Algorithm Based on Multi-View Synchronized Video Sequences},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={214-220},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013512800004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - 3D Human Body Model Reconstruction Algorithm Based on Multi-View Synchronized Video Sequences
SN - 978-989-758-754-2
AU - Wu Y.
PY - 2024
SP - 214
EP - 220
DO - 10.5220/0013512800004619
PB - SciTePress