Towards Reversible De-Identification in Video Sequences Using 3D Avatars and Steganography

Abstract

We propose a de-identification pipeline that protects the privacy of humans in video sequences by replacing them with rendered 3D human models, hence concealing their identity while retaining the naturalness of the scene. The original images of humans are steganographically encoded in the carrier image, i.e. the image containing the original scene and the rendered 3D human models. We qualitatively explore the feasibility of our approach, utilizing the Kinect sensor and its libraries to detect and localize human joints. A 3D avatar is rendered into the scene using the obtained joint positions, and the original human image is steganographically encoded in the new scene. Our qualitative evaluation shows reasonably good results that merit further exploration.

Files

Full Paper as PDF

BibTeX Citation

DOI

10.20532/ccvw.2015.0004

https://doi.org/10.20532/ccvw.2015.0004

BibTeX

@InProceedings{10.20532/ccvw.2015.0004,
  author =       {Martin Bla{\v z}evi{\' c} and Karla Brki{\' c} and
                  Tomislav Hrka{\' c}},
  title =        {Towards Reversible De-Identification in Video
                  Sequences Using 3D Avatars and Steganography},
  booktitle =    {Proceedings of the Croatian Compter Vision Workshop,
                  Year 3},
  pages =        {9-14},
  year =         2015,
  editor =       {Lon{\v c}ari{\' c}, Sven and Krapac, Josip},
  address =      {Zagreb},
  month =        {September},
  organization = {Center of Excellence for Computer Vision},
  publisher =    {University of Zagreb},
  abstract =     {We propose a de-identification pipeline that
                  protects the privacy of humans in video sequences by
                  replacing them with rendered 3D human models, hence
                  concealing their identity while retaining the
                  naturalness of the scene. The original images of
                  humans are steganographically encoded in the carrier
                  image, i{.}e{.}\ the image containing the original
                  scene and the rendered 3D human models. We
                  qualitatively explore the feasibility of our
                  approach, utilizing the Kinect sensor and its
                  libraries to detect and localize human joints. A 3D
                  avatar is rendered into the scene using the obtained
                  joint positions, and the original human image is
                  steganographically encoded in the new scene.  Our
                  qualitative evaluation shows reasonably good results
                  that merit further exploration.},
  doi =          {10.20532/ccvw.2015.0004},
  url =          {https://doi.org/10.20532/ccvw.2015.0004}
}