The repository is for the paper "VegFru: A Domain-Specific Dataset for Fine-grained Visual Categorization".
Download Links: [Paper] [Supplementary Material] [Project Page] [Download VegFru Dataset]
The VegFru dataset is available at here and this code is for the implementation of HybridNet.
./caffe-hybridnet: modified from Caffe (https://github.com/BVLC/caffe)
./hybridnet-dataset: the prototxt defining the models for each dataset (FGVC-Aircrafts)
cd caffe-hybridnet
sh data/jointaircrafts/model/v4_hyvggcpnet/ft_last_layer.sh
sh data/jointaircrafts/model/v4_hyvggcpnet/ft_all.sh
The pretrained models can be downloaded at here.
Please cite the following paper if you find this useful in your research:
@InProceedings{Hou2017VegFru,
Title = {VegFru: A Domain-Specific Dataset for Fine-grained Visual Categorization},
Author = {Saihui Hou, Yushan Feng and Zilei Wang},
Booktitle = {IEEE International Conference on Computer Vision},
Year = {2017}
}