FSANet: Frequency-Separated Attention Network for Image Super-Resolution

Daokuan Qu, Liulian Li and Rui Yao

Table of Contents

1. Introduction
2. Results
3. Preparation
4. Testing
5. Training
6. Citation

Introduction

The use of deep convolutional neural networks has significantly improved the performance of super-resolution. Employing deeper networks to enhance the non-linear mapping capability from low-resolution (LR) to high-resolution (HR) images has inadvertently weakened the information flow and disrupted long-term memory. Moreover, overly deep networks are challenging to train, thus failing to exhibit the expressive capability commensurate with their depth. High-frequency and low-frequency features in images play different roles in image super-resolution. Networks based on CNNs, which should focus more on high-frequency features, treat these two types of features equally. This results in redundant computations when processing low-frequency features and causes complex and detailed parts of the reconstructed images to appear as smooth as the background. To maintain long-term memory and focus more on the restoration of image details in networks with strong representational capabilities, we propose the Frequency-Separated Attention Network (FSANet), where dense connections ensure the full utilization of multi-level features. In the Feature Extraction Module (FEM), the use of the Res ASPP Module expands the network’s receptive field without increasing its depth. To differentiate between high-frequency and low-frequency features within the network, we introduce the Feature-Separated Attention Block (FSAB). Furthermore, to enhance the quality of the restored images using heuristic features, we incorporate attention mechanisms into the Low-Frequency Attention Block (LFAB) and the High-Frequency Attention Block (HFAB) for processing low-frequency and high-frequency features, respectively. The proposed network outperforms the current state-of-the-art methods in tests on benchmark datasets.

Results

Preparation

Requirements and Dependencies:

Here we list our used requirements and dependencies.

• Python: 3.11.5
• PyTorch: 2.2.0
• Torchvision: 0.17.0

Dataset：

Download the Div2k dataset at Div2k. Change data to .h5 file

python div2h5.py

Testing

Download the pre-trained model FSANet.

bash test.sh

Training

python train.py --patch_size 64 --batch_size 16 --lr 0.0001 --decay 150000  --scale 2 --gpu_index 0 --n_blocks 3
python train.py --patch_size 64 --batch_size 16 --lr 0.0001 --decay 150000  --scale 3 --gpu_index 2 --n_blocks 3 
python train.py --patch_size 64 --batch_size 16 --lr 0.0001 --decay 150000  --scale 4 --gpu_index 3 --n_blocks 3  

Citation

If you find the code useful in your research, please cite:

@article{qu2024frequency,
  title={Frequency-Separated Attention Network for Image Super-Resolution},
  author={Qu, Daokuan and Li, Liulian and Yao, Rui},
  journal={Applied Sciences},
  volume={14},
  number={10},
  pages={4238},
  year={2024},
  publisher={MDPI}
}

License

See MIT License