ML Anthology

Separating Appearance from Deformation

Nebojsa Jojic, Patrice Y. Simard, Brendan J. Frey, David Heckerman
ICCV 2001 pp. 288-294
doi:10.1109/ICCV.2001.937638 /iccv/2001/jojic2001iccv-separating/

Abstract

By representing images and image prototypes by linear subspaces spanned by "tangent vectors" (derivatives of an image with respect to translation, rotation, etc.), impressive invariance to known types of uniform distortion can be built into feedforward discriminators. We describe a new probability model that can jointly cluster data and learn mixtures of nonuniform, smooth deformation fields. Our fields are based on low-frequency wavelets, so they use very few parameters to model a wide range of smooth deformations (unlike, e.g., factor analysis, which uses a large number of parameters to model deformations). In spirit, our ideas are most similar to the idea of separating content from style published by Tenenbaum and Freeman. However, our models do not need labeled data for training, and thus allow for unsupervised separation of appearance from deformation. We give results on handwritten digit recognition and face recognition.

ICCV Semantic Scholar

Cite

Text

Jojic et al. "Separating Appearance from Deformation." IEEE/CVF International Conference on Computer Vision, 2001. doi:10.1109/ICCV.2001.937638

Markdown

[Jojic et al. "Separating Appearance from Deformation." IEEE/CVF International Conference on Computer Vision, 2001.](https://mlanthology.org/iccv/2001/jojic2001iccv-separating/) doi:10.1109/ICCV.2001.937638

BibTeX

@inproceedings{jojic2001iccv-separating,
  title     = {{Separating Appearance from Deformation}},
  author    = {Jojic, Nebojsa and Simard, Patrice Y. and Frey, Brendan J. and Heckerman, David},
  booktitle = {IEEE/CVF International Conference on Computer Vision},
  year      = {2001},
  pages     = {288-294},
  doi       = {10.1109/ICCV.2001.937638},
  url       = {https://mlanthology.org/iccv/2001/jojic2001iccv-separating/}
}