Fine-grained Visual Classification with High-temperature Refinement and Background Suppression

paper url: https://arxiv.org/abs/2303.06442

The high-temperature refinement module allows the model to learn the appropriate feature scales by refining the features map at different scales and improving the learning of diverse features. And, the background suppression module first splits the features map into foreground and background using classification confidence scores and suppresses feature values in low-confidence areas while enhancing discriminative features. The experimental results show that the proposed HERBS effectively fuses features of varying scales, suppresses background noise, discriminative features at appropriate scales for fine-grained visual classification.The proposed method achieves state-of-the-art performance on the CUB-200-2011 and NABirds benchmarks, surpassing 93% accuracy on both datasets. Thus, HERBS presents a promising solution for improving the performance of fine-grained visual classification tasks.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Please note that currently, only the Swin Transformer backbone is supported.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

1. Environment setting

// We move old version to ./v0/

1.0. Package

install requirements
replace folder timm/ to our timm/ folder (for ViT or Swin-T)

pytorch model implementation timm

recommand anaconda

recommand weights and biases

deepspeed // future works

1.1. Dataset

In this paper, we use 2 large bird's datasets to evaluate performance:

1.2. Our pretrained model

our pretrained model in
cub200: https://idocntnu-my.sharepoint.com/:f:/g/personal/81075001h_eduad_ntnu_edu_tw/EiXAfUuY3OJCjsF2cFeeEZoBKKKUDPLczUcJsAiRCjW1gg?e=LffbXM

1.3. OS

Windows10
Ubuntu20.04
macOS (CPU only)

2. Train

Single GPU Training
DataParallel (single machine multi-gpus)
DistributedDataParallel

(more information: https://pytorch.org/tutorials/intermediate/ddp_tutorial.html)

2.1. data

train data and test data structure:

├── tain/
│   ├── class1/
│   |   ├── img001.jpg
│   |   ├── img002.jpg
│   |   └── ....
│   ├── class2/
│   |   ├── img001.jpg
│   |   ├── img002.jpg
│   |   └── ....
│   └── ....
└──

2.2. configuration

you can directly modify yaml file (in ./configs/)

2.3. run

python main.py --c ./configs/config.yaml

model will save in ./records/{project_name}/{exp_name}/backup/

3. HeatMap

python vis_swin_l.py -pr ./backups/CUB200-2011/best.pt --image ./imgs/Laysan_Albatross_0098_621.jpg --save_name Laysan_Albatross_0098_621_heat

4. Evaluation

for CUB200-2011.

python run_evaluation.py -pr ./backups/CUB200-2011/ -ir ./datas/test/

-pr: pretrained root

├── pretrained_root/
│   |   ├── config.yaml
│   |   └── best.pt

-ir: images root

├── images_root/
│   |   ├── img001.jpg
│   |   ├── img002.jpg
│   |   └── ....

4.1 running

4.2 result

The results will show the fine-class accuracy and the number of outer errors. Outer errors refer to cases where the model predicts the wrong class and this class is not within the fine-class.

Future update

ResNet, Efficient support
Small model
Medical image support

chou141253/FGVC-HERBS