On the Exploitation of One Class Classification to Distinguish Food vs Non-Food Images

G. M. Farinella, D. Allegra, F. Stanco, S. Battiato

1st International Workshop on Multimedia Assisted Dietary Management, 2015

Paper Slides

CONTACT: 
Giovanni Maria Farinella
gfarinella@dmi.unict.it

  

The intent of this page is to provide the benchmark datasets together with training/testing details
such that algorithms for Food vs Non-Food classification can be properly compared.

Food vs Non-Food
DATASETS

The UNICT-FD889 food dataset used in the experiments for training and testing is available at this link.

The Food dataset (images downloaded from Flickr) used for testing purpose in the second experiment is available at this link.

The Non-Food dataset (images downloaded from Flickr) used for testing purpose in the second experiment is available at this link.


All the details on how to split the datasets to repeat the experiments can be found at this link.


Related works can be found at this link.