Camera Calibration and Shape Recovery from Videos of Two Mirrors

Abstract

This paper addresses the problem of motion and shape recovery from a two-mirror system which is able to generate five views of an object. Different from existing methods, this paper uses a short video instead of static snapshots so that it can help with action recognition once the 3D visual hull model is reconstructed. In order to solve the problem, this paper shows the geometry relationship between the two-mirror system and circular motion, so that the two-mirror system can be solved as circular motions. Different from the approach of Zhang et al. [ 22 ], we avoid using the vanishing point of X-axis which would cause accumulate error when calculating the epipoles of two views. Results of comparative experiments and the 3D visual hull of model show the feasibility and the accuracy of the proposed approach.

Cite

Text

Chen and Zhang. "Camera Calibration and Shape Recovery from Videos of Two Mirrors." European Conference on Computer Vision Workshops, 2014. doi:10.1007/978-3-319-16178-5_53

Markdown

[Chen and Zhang. "Camera Calibration and Shape Recovery from Videos of Two Mirrors." European Conference on Computer Vision Workshops, 2014.](https://mlanthology.org/eccvw/2014/chen2014eccvw-camera/) doi:10.1007/978-3-319-16178-5_53

BibTeX

@inproceedings{chen2014eccvw-camera,
  title     = {{Camera Calibration and Shape Recovery from Videos of Two Mirrors}},
  author    = {Chen, Quanxin and Zhang, Hui},
  booktitle = {European Conference on Computer Vision Workshops},
  year      = {2014},
  pages     = {757-768},
  doi       = {10.1007/978-3-319-16178-5_53},
  url       = {https://mlanthology.org/eccvw/2014/chen2014eccvw-camera/}
}