TITLE
Mobile Human Shape Superimposition: An Initial Approach using OpenPose
ABSTRACT
When it comes to visitors at museums and heritage attractions objects speak for themselves. Nevertheless, it is important to give those visitors the best experience possible as, this will lead to an increase in the visits number, enhance the perception and value of an organisation, and boost the sales. With the aim of enhancing a traditional museum visit, a mobile Augmented Reality (AR) framework is being developed as part of the Mobile Five Senses Augmented Reality (M5SAR) system for museums project. This paper presents an initial approach to human shape detection and AR content superimposition, achieved by combining information of human body joints with shape segmentation or texture overlapping. The OpenPose model was used to compute the body joints and the GrabCut algorithm for person segmentation, allowing to fit clothes segments to persons moving in real environments. The initial results and proof-of-concept are shown.
KEYWORDS
Pose Estimation, Segmentation, Human Shape Superimposition, Convolutional Neural Networks
Cite this paper
Roman Bajireanu, Joao A. R. Pereira, Ricardo J. M. Veiga, Joao D. P. Sardo, Pedro J. S. Cardoso, Roberto Lam, Joao M. F. Rodrigues. (2019) Mobile Human Shape Superimposition: An Initial Approach using OpenPose. International Journal of Computers, 4, 1-8
Copyright © 2018 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0