Dreaming More Data: Class-dependent Distributions over Diffeomorphisms for Learned Data Augmentation

@inproceedings{Hauberg2016DreamingMD,
  title={Dreaming More Data: Class-dependent Distributions over Diffeomorphisms for Learned Data Augmentation},
  author={S\oren Hauberg and Oren Freifeld and Anders Boesen Lindbo Larsen and John W. Fisher and Lars Kai Hansen},
  booktitle={AISTATS},
  year={2016}
}
Data augmentation is a key element in training highdimensional models. In this approach, one synthesizes new observations by applying pre-specified transformations to the original training data; e.g. new images are formed by rotating old ones. Current augmentation schemes, however, rely on manual specification of the applied transformations, making data augmentation an implicit form of feature engineering. Working towards true end-to-end learning, we suggest to learn the applied transformations… CONTINUE READING

6 Figures & Tables

Topics

Statistics

01020201620172018
Citations per Year

Citation Velocity: 12

Averaging 12 citations per year over the last 3 years.

Learn more about how we calculate this metric in our FAQ.
  • GitHub repos referencing this paper

    • AIDE

  • Presentations referencing similar topics