Food-101 – Automatic point-of-interest image cropping via ensembled convolutionalization

Andrea Asperti, Pietro Battilana

This page contains the architecure (json) and weights (hdf5) for the discriminative networks trained with transfer learning for the FOOD-101 dataset. We also gives the respective accuracies on the validation set.

Please, refer to the article for a description of the nets and their training process.

Back to home