Fine-Grained Image Classification on Stanford Cars (FGVC-Stanford Cars)

Setup

Setup anaconda environment using environment.yml file.

conda env create --name DiffuseMix --file=environment.yml
conda remove -n DiffuseMix --all # In case environment installation faileds

Dataset Structure

train
 └─── class 1
          └───── n04355338_22023.jpg
 └─── ...

val
 └─── class 1
          └───── n03786901_5410.jpg
 └─── ...

Train Examples

To introduce the structural complexity, you can download fractal image dataset from here Fractal Dataset

`python3 main.py --train_dir PATH --fractal_dir PATH --prompts sunset,Autumn

Comparison with SOTA Mixup

Method	Stanford Cars
Vanilla(CVPR'16)	85.52
RA(NIPS'20)	87.79
AdaAug(ICLR'22)	88.49
PuzzleMix(ICML'20)	89.68
Co-Mixup(ICLR'21)	89.53
Guided-AP(AAAI'23)	90.27
DiffuseMix	91.26

Test on Validation Set

Accuracy of the network on the 8041 test images: 91.23%

Citation

If you find our work useful in your research please consider citing our paper:

@article{diffuseMix2024,
  title={DIFFUSEMIX: Label-Preserving Data Augmentation with Diffusion Models},
  author={Khawar Islam, Muhammad Zaigham Zaheer, Arif Mahmood, Karthik Nandakumar},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2024}
}