Easy and Fast Reconstruction of a 3D Avatar with an RGB-D Sensor.

Aihua Mao,Hong Zhang,Guoqiang Han,Yuxin Liu,Guiqing Li,Yinglong Zheng

doi:10.3390/s17051113

Abstract

This paper proposes a new easy and fast 3D avatar reconstruction method using an RGB-D sensor. Users can easily implement human body scanning and modeling just with a personal computer and a single RGB-D sensor such as a Microsoft Kinect within a small workspace in their home or office. To make the reconstruction of 3D avatars easy and fast, a new data capture strategy is proposed for efficient human body scanning, which captures only 18 frames from six views with a close scanning distance to fully cover the body; meanwhile, efficient alignment algorithms are presented to locally align the data frames in the single view and then globally align them in multi-views based on pairwise correspondence. In this method, we do not adopt shape priors or subdivision tools to synthesize the model, which helps to reduce modeling complexity. Experimental results indicate that this method can obtain accurate reconstructed 3D avatar models, and the running performance is faster than that of similar work. This research offers a useful tool for the manufacturers to quickly and economically create 3D avatars for products design, entertainment and online shopping.

Highlights

Modeling real humans in the virtual world is always a popular research topic since virtual avatars are essential for products design and manufacturing for living and have a great deal of applications for games, movies, video conferences and online shopping
Considering alignment efficiency, we propose a Another problem in multi-view alignment is non-rigid deformation, which may be caused by three-step multi-view alignment algorithm for the sequence of point clouds in different views
We demonstrate the reconstruction results for 3D avatars based on our method in different scanning and compare them with the results similar works

Summary

Introduction

Modeling real humans in the virtual world is always a popular research topic since virtual avatars are essential for products design and manufacturing for living and have a great deal of applications for games, movies, video conferences and online shopping. Based on the RGB-D data from Kinect, there has been a great deal of research on human reconstruction to achieve cost savings and greater efficiency [4,5,6,7,8,9], the final reconstructed human model is still far away to be acceptable. The existing research is primarily concerned with the following: (1) the use of fewer Kinects to increase modeling convenience and save on costs, as in [4,6,7,9]; (2) the use of fewer shape priors to reduce reconstruction complexity and obtain more of the actual body silhouette, such as in [5,8]; (3) the development of more sophisticated alignment algorithms to improve modeling.

Results

Discussion

Conclusion