ML Anthology

Are Sparse Representations Really Relevant for Image Classification?

Roberto Rigamonti, Matthew A. Brown, Vincent Lepetit
CVPR 2011 pp. 1545-1552
doi:10.1109/CVPR.2011.5995313 /cvpr/2011/rigamonti2011cvpr-sparse/

Abstract

Recent years have seen an increasing interest in sparse representations for image classification and object recognition, probably motivated by evidence from the analysis of the primate visual cortex. It is still unclear, however, whether or not sparsity helps classification. In this paper we evaluate its impact on the recognition rate using a shallow modular architecture, adopting both standard filter banks and filter banks learned in an unsupervised way. In our experiments on the CIFAR-10 and on the Caltech-101 datasets, enforcing sparsity constraints actually does not improve recognition performance. This has an important practical impact in image descriptor design, as enforcing these constraints can have a heavy computational cost.

PDF CVPR Semantic Scholar

Cite

Text

Rigamonti et al. "Are Sparse Representations Really Relevant for Image Classification?." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2011. doi:10.1109/CVPR.2011.5995313

Markdown

[Rigamonti et al. "Are Sparse Representations Really Relevant for Image Classification?." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2011.](https://mlanthology.org/cvpr/2011/rigamonti2011cvpr-sparse/) doi:10.1109/CVPR.2011.5995313

BibTeX

@inproceedings{rigamonti2011cvpr-sparse,
  title     = {{Are Sparse Representations Really Relevant for Image Classification?}},
  author    = {Rigamonti, Roberto and Brown, Matthew A. and Lepetit, Vincent},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2011},
  pages     = {1545-1552},
  doi       = {10.1109/CVPR.2011.5995313},
  url       = {https://mlanthology.org/cvpr/2011/rigamonti2011cvpr-sparse/}
}