File size: 8,346 Bytes
7490281 84f4c1d 0a0b275 7490281 0a0b275 7490281 0a0b275 7490281 0a0b275 7490281 0a0b275 7490281 0a0b275 7490281 76fd501 7490281 0a0b275 7490281 0a0b275 7490281 0a0b275 7490281 4d2c08a a4099c3 7490281 4d2c08a a4099c3 7490281 4d2c08a a4099c3 7490281 4d2c08a a4099c3 7490281 0a0b275 7490281 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 |
---
license: apache-2.0
tags:
- super-image
- image-super-resolution
datasets:
- eugenesiow/Div2k
- eugenesiow/Set5
- eugenesiow/Set14
- eugenesiow/BSD100
- eugenesiow/Urban100
metrics:
- pnsr
- ssim
---
# Attention in Attention Network for Image Super-Resolution (A2N)
A2N model pre-trained on DIV2K (800 images training, augmented to 4000 images, 100 images validation) for 2x, 3x and 4x image super resolution. It was introduced in the paper [Attention in Attention Network for Image Super-Resolution](https://arxiv.org/abs/2104.09497) by Chen et al. (2021) and first released in [this repository](https://github.com/haoyuc/A2N).
The goal of image super resolution is to restore a high resolution (HR) image from a single low resolution (LR) image. The image below shows the ground truth (HR), the bicubic upscaling x2 and model upscaling x2.
## Model description
The A2N model proposes an attention in attention network (A2N) for highly accurate image SR. Specifically, the A2N consists of a non-attention branch and a coupling attention branch. Attention dropout module is proposed to generate dynamic attention weights for these two branches based on input features that can suppress unwanted attention adjustments. This allows attention modules to specialize to beneficial examples without otherwise penalties and thus greatly improve the capacity of the attention network with little parameter overhead.
More importantly the model is lightweight and fast to train (~1.5m parameters, ~4mb).
## Intended uses & limitations
You can use the pre-trained models for upscaling your images 2x, 3x and 4x. You can also use the trainer to train a model on your own dataset.
### How to use
The model can be used with the [super_image](https://github.com/eugenesiow/super-image) library:
```bash
pip install super-image
```
Here is how to use a pre-trained model to upscale your image:
```python
from super_image import A2nModel, ImageLoader
from PIL import Image
import requests
url = 'https://paperswithcode.com/media/datasets/Set5-0000002728-07a9793f_zA3bDjj.jpg'
image = Image.open(requests.get(url, stream=True).raw)
model = A2nModel.from_pretrained('eugenesiow/a2n', scale=2) # scale 2, 3 and 4 models available
inputs = ImageLoader.load_image(image)
preds = model(inputs)
ImageLoader.save_image(preds, './scaled_2x.png') # save the output 2x scaled image to `./scaled_2x.png`
ImageLoader.save_compare(inputs, preds, './scaled_2x_compare.png') # save an output comparing the super-image with a bicubic scaling
```
[](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Upscale_Images_with_Pretrained_super_image_Models.ipynb "Open in Colab")
## Training data
The models for 2x, 3x and 4x image super resolution were pretrained on [DIV2K](https://huggingface.co/datasets/eugenesiow/Div2k), a dataset of 800 high-quality (2K resolution) images for training, augmented to 4000 images and uses a dev set of 100 validation images (images numbered 801 to 900).
## Training procedure
### Preprocessing
We follow the pre-processing and training method of [Wang et al.](https://arxiv.org/abs/2104.07566).
Low Resolution (LR) images are created by using bicubic interpolation as the resizing method to reduce the size of the High Resolution (HR) images by x2, x3 and x4 times.
During training, RGB patches with size of 64×64 from the LR input are used together with their corresponding HR patches.
Data augmentation is applied to the training set in the pre-processing stage where five images are created from the four corners and center of the original image.
We need the huggingface [datasets](https://huggingface.co/datasets?filter=task_ids:other-other-image-super-resolution) library to download the data:
```bash
pip install datasets
```
The following code gets the data and preprocesses/augments the data.
```python
from datasets import load_dataset
from super_image.data import EvalDataset, TrainDataset, augment_five_crop
augmented_dataset = load_dataset('eugenesiow/Div2k', 'bicubic_x4', split='train')\
.map(augment_five_crop, batched=True, desc="Augmenting Dataset") # download and augment the data with the five_crop method
train_dataset = TrainDataset(augmented_dataset) # prepare the train dataset for loading PyTorch DataLoader
eval_dataset = EvalDataset(load_dataset('eugenesiow/Div2k', 'bicubic_x4', split='validation')) # prepare the eval dataset for the PyTorch DataLoader
```
### Pretraining
The model was trained on GPU. The training code is provided below:
```python
from super_image import Trainer, TrainingArguments, A2nModel, A2nConfig
training_args = TrainingArguments(
output_dir='./results', # output directory
num_train_epochs=1000, # total number of training epochs
)
config = A2nConfig(
scale=4, # train a model to upscale 4x
)
model = A2nModel(config)
trainer = Trainer(
model=model, # the instantiated model to be trained
args=training_args, # training arguments, defined above
train_dataset=train_dataset, # training dataset
eval_dataset=eval_dataset # evaluation dataset
)
trainer.train()
```
[](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Train_super_image_Models.ipynb "Open in Colab")
## Evaluation results
The evaluation metrics include [PSNR](https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio#Quality_estimation_with_PSNR) and [SSIM](https://en.wikipedia.org/wiki/Structural_similarity#Algorithm).
Evaluation datasets include:
- Set5 - [Bevilacqua et al. (2012)](https://huggingface.co/datasets/eugenesiow/Set5)
- Set14 - [Zeyde et al. (2010)](https://huggingface.co/datasets/eugenesiow/Set14)
- BSD100 - [Martin et al. (2001)](https://huggingface.co/datasets/eugenesiow/BSD100)
- Urban100 - [Huang et al. (2015)](https://huggingface.co/datasets/eugenesiow/Urban100)
The results columns below are represented below as `PSNR/SSIM`. They are compared against a Bicubic baseline.
|Dataset |Scale |Bicubic |A2N |
|--- |--- |--- |--- |
|Set5 |2x |33.64/0.9292 |**37.87/0.9602** |
|Set5 |3x |30.39/0.8678 |**34.8/0.9387** |
|Set5 |4x |28.42/0.8101 |**32.07/0.8933** |
|Set14 |2x |30.22/0.8683 |**33.45/0.9162** |
|Set14 |3x |27.53/0.7737 |**30.94/0.8568** |
|Set14 |4x |25.99/0.7023 |**28.56/0.7801** |
|BSD100 |2x |29.55/0.8425 |**32.11/0.8987** |
|BSD100 |3x |27.20/0.7382 |**29.56/0.8173** |
|BSD100 |4x |25.96/0.6672 |**27.54/0.7342** |
|Urban100 |2x |26.66/0.8408 |**31.71/0.9240** |
|Urban100 |3x | |**28.95/0.8671** |
|Urban100 |4x |23.14/0.6573 |**25.89/0.7787** |
You can find a notebook to easily run evaluation on pretrained models below:
[](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Evaluate_Pretrained_super_image_Models.ipynb "Open in Colab")
## BibTeX entry and citation info
```bibtex
@misc{chen2021attention,
title={Attention in Attention Network for Image Super-Resolution},
author={Haoyu Chen and Jinjin Gu and Zhi Zhang},
year={2021},
eprint={2104.09497},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
``` |