Handling Uncertain Tags in Visual Recognition

Abstract

Gathering accurate training data for recognizing a set of attributes or tags on images or videos is a challenge. Obtaining labels via manual effort or from weakly-supervised data typically results in noisy training labels. We develop the FlipSVM, a novel algorithm for handling these noisy, structured labels. The FlipSVM models label noise by "flipping" labels on training examples. We show empirically that the FlipSVM is effective on images-and-attributes and video tagging datasets.

Cite

Text

Vahdat and Mori. "Handling Uncertain Tags in Visual Recognition." International Conference on Computer Vision, 2013. doi:10.1109/ICCV.2013.462

Markdown

[Vahdat and Mori. "Handling Uncertain Tags in Visual Recognition." International Conference on Computer Vision, 2013.](https://mlanthology.org/iccv/2013/vahdat2013iccv-handling/) doi:10.1109/ICCV.2013.462

BibTeX

@inproceedings{vahdat2013iccv-handling,
  title     = {{Handling Uncertain Tags in Visual Recognition}},
  author    = {Vahdat, Arash and Mori, Greg},
  booktitle = {International Conference on Computer Vision},
  year      = {2013},
  doi       = {10.1109/ICCV.2013.462},
  url       = {https://mlanthology.org/iccv/2013/vahdat2013iccv-handling/}
}