X-TRON404/MAN

Multi-scale Attention Network for Single Image Super-Resolution

MAN

Codes for paper "Multi-scale Attention Network for Single Image Super-Resolution".

@article{wang2022multi,
  title={Multi-scale Attention Network for Single Image Super-Resolution},
  author={Wang, Yan and Li, Yusen and Wang, Gang and Liu, Xiaoguang},
  journal={arXiv preprint arXiv:2209.14145},
  year={2022}
}

Implementary Details

Network architecture: MAB number (n_resblocks): 5/24/36, channel width (n_feats): 48/60/180 for tiny/light/base MAN.

Overview of the proposed MAN constituted of three components: the shallow feature extraction module (SF), the deep feature extraction module (DF) based on multiple multi-scale attention blocks (MAB), and the high-quality image reconstruction module.

 

Component details: three multi-scale decomposition modes are utilized in MLKA. The 7×7 depth-wise convolution is used in the GSAU.

Details of Multi-scale Large Kernel Attention (MLKA), Gated Spatial Attention Unit (GSAU), and Large Kernel Attention Tail (LKAT).

 

Results and Models

Pretrained models avialable at Google Drive and Baidu Pan (pwd: mans for all links).

Results of our MAN-tiny/light/base models. Set5 validation set is used below to show the general performance. The visual results of five testsets are provided in the last column.

Methods	Params	Madds	PSNR/SSIM (x2)	PSNR/SSIM (x3)	PSNR/SSIM (x4)	Results
MAN-tiny	150K	8.4G	37.91/0.9603	-	32.07/0.8930	x2/x4
MAN-light	840K	47.1G	38.18/0.9612	34.65/0.9292	32.50/0.8988	x2/x3/x4
MAN+	8712K	495G	38.44/0.9623	34.97/0.9315	32.87/0.9030	x2/x3/x4

Training and Testing

The BasicSR framework is utilized to train our MAN, also testing.

Acknowledgements

We would thank VAN and BasicSR for their enlightening work!