On Sparsity and Overcompleteness in Image Models
Abstract
Computational models of visual cortex, and in particular those based on sparse coding, have enjoyed much recent attention. Despite this currency, the question of how sparse or how over-complete a sparse representation should be, has gone without principled answer. Here, we use Bayesian model-selection methods to address these questions for a sparse-coding model based on a Student-t prior. Having validated our methods on toy data, we find that natural images are indeed best modelled by extremely sparse distributions; although for the Student-t prior, the associated optimal basis size is only modestly overcomplete.
Cite
Text
Berkes et al. "On Sparsity and Overcompleteness in Image Models." Neural Information Processing Systems, 2007.Markdown
[Berkes et al. "On Sparsity and Overcompleteness in Image Models." Neural Information Processing Systems, 2007.](https://mlanthology.org/neurips/2007/berkes2007neurips-sparsity/)BibTeX
@inproceedings{berkes2007neurips-sparsity,
title = {{On Sparsity and Overcompleteness in Image Models}},
author = {Berkes, Pietro and Turner, Richard and Sahani, Maneesh},
booktitle = {Neural Information Processing Systems},
year = {2007},
pages = {89-96},
url = {https://mlanthology.org/neurips/2007/berkes2007neurips-sparsity/}
}