Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Algolzw/NCNet
[AIM & ECCVW 2022] Fast Nearest Convolution for Real-Time Image Super-Resolution
https://github.com/Algolzw/NCNet
aim2022 eccv2022 nearest-convolution real-time super-resolution tensorflow2
Last synced: about 2 months ago
JSON representation
[AIM & ECCVW 2022] Fast Nearest Convolution for Real-Time Image Super-Resolution
- Host: GitHub
- URL: https://github.com/Algolzw/NCNet
- Owner: Algolzw
- License: apache-2.0
- Created: 2022-08-20T04:27:48.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2023-03-23T13:38:37.000Z (over 1 year ago)
- Last Synced: 2024-07-21T21:41:42.906Z (2 months ago)
- Topics: aim2022, eccv2022, nearest-convolution, real-time, super-resolution, tensorflow2
- Language: Python
- Homepage:
- Size: 2.99 MB
- Stars: 71
- Watchers: 2
- Forks: 3
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## Fast Nearest Convolution for Real-Time Image Super-Resolution, AIM & ECCV Workshops 2022, [[Paper]](https://arxiv.org/abs/2208.11609)
![ts](figs/ts.png)
### Update
[**2022.11.14**] A more flexible implementation of the nearest convolution initialization is provided in [this issue](https://github.com/Algolzw/NCNet/issues/5).[**2022.11.12**] We provide a simple implementation of [NCNet model](https://github.com/Algolzw/NCNet/blob/main/torch_code/ncnet.py) using PyTorch in this [torch_code](https://github.com/Algolzw/NCNet/tree/main/torch_code) directory.
[**2022.08.25**] We have uploaded the pretrained model in Releases as [V1.0](https://github.com/Algolzw/NCNet/releases/tag/V1.0)
### Dependencies
- OS: Ubuntu 18.04
- Python: Python 3.7
- Tensorflow 2.9.1
- nvidia :
- cuda: 11.2
- cudnn: 8.1.0
- Other reference requirements### Performance of our Nearest Convolution
![speed](figs/speed.png)
| Upsample methods | CPU | GPU | NPU | PSNR |
| ---- | ---- | ---- | ---- | ---- |
| nearest | 23.1ms | **19.0ms** | 55.0ms | 26.67 |
| bilinear | 77.7ms | 21.0ms | 128.2ms | **27.67** |
| Conv3+depth2space | 30.8ms | 26.5ms | 43.8ms | - |
| NearestConv+depth2space | **15.9ms** | 20.3ms | **14.8ms** | 26.67 |### Model Training
```python3
python main.py
```
Then the trained keras model will be saved in ```ckpt/basenet/model``` folder.### Model Validation
```python3
python eval.py
```
Then the results of original keras model will be saved in ```original_output``` folder and you can calculate the validation PSNR by run ```python calculate_PSNR.py```### Convert to TFLite
``` bash
python generate_tflite.py
```
Then the converted tflite model will be saved in ```TFLite ``` folder.### TFLite Model Validation
``` bash
python test_tflite.py
```
Then the results of TFLite model will be saved in ```results ``` folder.### Other Details
* The input image range is [0, 255].
* Number of parameters: 52,279 (53K)
* Average PSNR on DIV2K validation data: 30.27 dB
* Training data: DIV2K.## Citations
If this repo helps your research or work, please consider citing our work.
The following is a BibTeX reference.```
@inproceedings{luo2023fast,
title={Fast nearest convolution for real-time efficient image super-resolution},
author={Luo, Ziwei and Li, Youwei and Yu, Lei and Wu, Qi and Wen, Zhihong and Fan, Haoqiang and Liu, Shuaicheng},
booktitle={Computer Vision--ECCV 2022 Workshops: Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part II},
pages={561--572},
year={2023},
organization={Springer}
}
```## Contact
email: [[email protected]]