Object Recognition by Scene Alignment

Abstract

Current object recognition systems can only recognize a limited number of object categories; scaling up to many categories is the next challenge. We seek to build a system to recognize and localize many different object categories in complex scenes. We achieve this through a simple approach: by matching the input im- age, in an appropriate representation, to images in a large training set of labeled images. Due to regularities in object identities across similar scenes, the retrieved matches provide hypotheses for object identities and locations. We build a prob- abilistic model to transfer the labels from the retrieval set to the input image. We demonstrate the effectiveness of this approach and study algorithm component contributions using held-out test sets from the LabelMe database.

Cite

Text

Russell et al. "Object Recognition by Scene Alignment." Neural Information Processing Systems, 2007.

Markdown

[Russell et al. "Object Recognition by Scene Alignment." Neural Information Processing Systems, 2007.](https://mlanthology.org/neurips/2007/russell2007neurips-object/)

BibTeX

@inproceedings{russell2007neurips-object,
  title     = {{Object Recognition by Scene Alignment}},
  author    = {Russell, Bryan and Torralba, Antonio and Liu, Ce and Fergus, Rob and Freeman, William T.},
  booktitle = {Neural Information Processing Systems},
  year      = {2007},
  pages     = {1241-1248},
  url       = {https://mlanthology.org/neurips/2007/russell2007neurips-object/}
}