hongbo-sun / Gard

Code for Graph-based High-Order Relation Discovery for Fine-grained Recognition in CVPR 2021

Introduction

This code provides an initial version for the implementation of the CVPR2021 paper "Graph-based High-Order Relation Discovery for Fine-grained Recognition" Paper Link. The projects are still under construction.

For our projects, please refer to http://cvteam.net/projects/2021/Gard/Gard.html

Using API

Please refer to /doc/build/html/index.html, including function definition.

or finding at: http://cvteam.net/projects/2021/Gard/html/

How to run

For Training:

1. Download the benchmark dataset and unzip them in your customized path.

   CUB-200-2011 http://www.vision.caltech.edu/visipedia/CUB-200-2011.html

   FGVC-Aircraft https://www.robots.ox.ac.uk/~vgg/data/fgvc-aircraft/

   Stanford-Cars http://ai.stanford.edu/~jkrause/cars/car_dataset.html

   NABirds https://dl.allaboutbirds.org/nabirds

2. Build the Train/validation partition by yourself or download the files from here.

3. Modify the configuration files in /config/config.py and /config/default.py

4. Dataset config

   4-1 Set the dataset path in /config/config.py if using CUB dataset

   4-2 Set the dataset path in /datasets/UnifiedLoader.py if using other datasets

   4-3 Add functions and make your own datasets in /datasets/UnifiedLoader.py

5. Modify Line 70~76, uncomment used dataset and comment out other datasets

6. Run train.py for training.

For Quick Testing:

1. repeat or confirm the operations in training 1~5

2. Modify the class_num in /config/config.py

3. Download Testing weights.

   For example, CUB-200-2011 Model Link

4. Modify the Test weights dir in /config/config.py

5. Run test.py

Prerequisites

PyTorch, tqdm, torchvsion, Pillow, cv2

Running on Two GPUs to achieve the reported performance.

If on other GPU settings, the hyper params should be modified to achieve similar results.

Known issues

The performance would be fluctuated by 0.1~0.5% in different GPUs and PyTorch platforms. Pytorch versions higher than 1.3.1 are tested. Some results are much higher than the reported results (e.g. results on CUB dataset can reach 90.4% vs reported 89.6% when changing different platforms).

To do

1. The project is still ongoing, finding suitable platforms and GPU devices for complete stable results.

2. The project is re-constructed for better understanding, we release this version for a quick preview of our paper.

Citations:

Please remember to cite us if u find this useful.

@inproceedings{zhao2021graph,
  title={Graph-based high-order relation discovery for fine-grained recognition},
  author={Zhao, Yifan and Yan, Ke and Huang, Feiyue and Li, Jia},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={15079--15088},
  year={2021}
}