Boosting Few-shot Fine-grained Recognition with Background Suppression and Foreground Alignment (BSFA-FSFG)

Authors: Zican Zha, Hao Tang, Yunlian Sun, and Jinhui Tang.

This repository provides code for "Boosting Few-shot Fine-grained Recognition with Background Suppression and Foreground Alignment" IEEE TCSVT 2023.

Requirements

• Python 3.6
• Pytorch >= 1.7.0
• Torchvision = 0.10
• scikit-image = 0.18.1

Data Preparation

Download Datasets from Baidu Drive (extraction code: ZZC3)
Download Datasets from Google Drive

How to run

python train.py --dataset [type of dataset] --model [backbone] --num_classes [num-classes] --nExemplars [num-shots]
python test.py --dataset CUB-200-2011 --model R --num_classes 100 --nExemplars 5

# Example: run on CUB dataset, ResNet-12 backbone, 5-way 1-shot
python train.py --dataset CUB-200-2011 --model R --num_classes 100 --nExemplars 1
python test.py --dataset CUB-200-2011 --model R --num_classes 100 --nExemplars 1

Citation

Please cite our paper if you find the work useful, thanks!

   @article{zha2023boosting,
     title={Boosting Few-shot Fine-grained Recognition with Background Suppression and Foreground Alignment},
     author={Zha, Zican and Tang, Hao and Sun, Yunlian and Tang, Jinhui},
     journal={IEEE Transactions on Circuits and Systems for Video Technology},
     year={2023},
     publisher={IEEE},
     doi={10.1109/TCSVT.2023.3236636}
  }

Other related papers

   @article{TangYLT22,
        author    = {Hao Tang and Chengcheng Yuan and Zechao Li and Jinhui Tang},
        title     = {Learning attention-guided pyramidal features for few-shot fine-grained recognition},
        journal   = {Pattern Recognit.},
        volume    = {130},
        pages     = {108792},
        year      = {2022}
   }

Acknowledgement

This code is based on the implementations of fewshot-CAN.