ML Anthology

Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection

David Novotny, Samuel Albanie, Diane Larlus, Andrea Vedaldi
CVPR 2018
doi:10.1109/CVPR.2018.00383 /cvpr/2018/novotny2018cvpr-selfsupervised/

Abstract

Self-supervision can dramatically cut back the amount of manually-labelled data required to train deep neural networks. While self-supervision has usually been considered for tasks such as image classification, in this paper we aim at extending it to geometry-oriented tasks such as semantic matching and part detection. We do so by building on several recent ideas in unsupervised landmark detection. Our approach learns dense distinctive visual descriptors from an unlabeled dataset of images using synthetic image transformations. It does so by means of a robust probabilistic formulation that can introspectively determine which image regions are likely to result in stable image matching. We show empirically that a network pre-trained in this manner requires significantly less supervision to learn semantic object parts compared to numerous pre-training alternatives. We also show that the pre-trained representation is excellent for semantic object matching.

PDF CVPR Semantic Scholar

Cite

Text

Novotny et al. "Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018. doi:10.1109/CVPR.2018.00383

Markdown

[Novotny et al. "Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018.](https://mlanthology.org/cvpr/2018/novotny2018cvpr-selfsupervised/) doi:10.1109/CVPR.2018.00383

BibTeX

@inproceedings{novotny2018cvpr-selfsupervised,
  title     = {{Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection}},
  author    = {Novotny, David and Albanie, Samuel and Larlus, Diane and Vedaldi, Andrea},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2018},
  doi       = {10.1109/CVPR.2018.00383},
  url       = {https://mlanthology.org/cvpr/2018/novotny2018cvpr-selfsupervised/}
}