Laser & Optoelectronics Progress, Volume. 62, Issue 16, 1637002(2025)
Lightweight Image Super-Resolution Reconstruction Incorporating Dual-Stream Feature Enhancement
Figures & Tables(13)
Fig. 8. Comparison of reconstruction effects of images at ×4 scale in Urban100 dataset with/without Dysample. (a) With Dysample; (b) without Dysample
Table 1. Peformances of different algorithms on each dataset
View tableTable 1. Peformances of different algorithms on each dataset
Method Scale Set5 Set14 BSD100 Urban100 Manga109 PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM Bicubic ×2 33.66 0.9299 30.24 0.8688 29.56 0.8431 26.88 0.8403 30.80 0.9339 SRCNN 36.66 0.9542 32.45 0.9067 31.36 0.8879 29.50 0.8946 35.60 0.9663 VDSR 37.53 0.9587 33.03 0.9124 31.90 0.8960 30.76 0.9140 37.22 0.9750 DRCN 37.63 0.9588 33.04 0.9118 31.85 0.8942 30.75 0.9133 37.55 0.9732 CARN 37.76 0.9590 33.52 0.9166 32.09 0.8978 31.92 0.9256 38.36 0.9765 IMDN 38.00 0.9605 33.63 0.9177 32.19 0.8996 32.17 0.9283 38.88 0.9774 EDSR 37.99 0.9604 33.57 0.9175 32.16 0.8994 31.98 0.9272 38.54 0.9769 SMSR 38.00 0.9601 33.64 0.9179 32.17 0.8990 32.19 0.9284 38.76 0.9771 RFDN 38.05 0.9606 33.68 0.9184 32.16 0.8994 32.12 0.9278 38.88 0.9773 SAFMN 38.00 0.9605 33.54 0.9177 32.16 0.8995 31.84 0.9256 38.71 0.9771 ESRT 38.03 0.9600 33.75 0.9184 32.25 0.9001 32.58 0.9318 39.12 0.9774 HNCT 38.08 0.9608 33.65 0.9182 32.22 0.9001 32.22 0.9294 38.87 0.9774 LBNet 38.05 0.9607 33.65 0.9177 32.16 0.8994 32.30 0.9291 38.88 0.9775 NGswin 38.10 0.9610 33.79 0.9199 32.27 0.9008 32.53 0.9324 38.97 0.9777 SwinIR-Light 38.14 0.9611 33.86 0.9206 32.31 0.9012 32.76 0.9340 39.12 0.9783 DSLSR 38.13 0.9610 33.94 0.9211 32.31 0.9016 32.76 0.9340 39.13 0.9779 Bicubic ×3 30.39 0.8682 27.55 0.7742 27.21 0.7385 24.46 0.7349 26.96 0.8556 SRCNN 32.75 0.9090 29.28 0.8209 28.41 0.7863 26.24 0.7989 32.01 0.9340 VDSR 33.66 0.9213 29.77 0.8314 28.82 0.7976 27.14 0.8279 32.01 0.9340 DRCN 33.82 0.9226 29.76 0.8311 28.80 0.7963 27.15 0.8276 32.24 0.9343 CARN 34.29 0.9255 30.29 0.8407 29.06 0.8034 28.06 0.8493 33.50 0.9440 IMDN 34.36 0.9270 30.32 0.8417 29.09 0.8046 28.17 0.8519 33.61 0.9445 EDSR 34.37 0.9270 30.28 0.8417 29.09 0.8052 28.15 0.8527 33.45 0.9439 SMSR 34.40 0.9270 30.33 0.8412 29.10 0.8050 28.25 0.8536 33.68 0.9445 RFDN 34.41 0.9273 30.34 0.8420 29.09 0.8050 28.21 0.8525 33.67 0.9449 SAFMN 34.34 0.9267 30.33 0.8418 29.08 0.8048 27.95 0.8474 33.52 0.9437 ESRT 34.42 0.9268 30.43 0.8433 29.15 0.8063 28.46 0.8574 33.95 0.9455 HNCT 34.47 0.9275 30.44 0.8439 29.15 0.8067 28.28 0.8557 33.81 0.9459 LBNet 34.47 0.9277 30.38 0.8417 29.13 0.8061 28.42 0.8559 33.82 0.9406 NGswin 34.52 0.9282 30.53 0.8456 29.19 0.8078 28.52 0.8603 33.89 0.9470 SwinIR-Light 34.62 0.9289 30.54 0.8463 29.20 0.8082 28.66 0.8624 33.98 0.9478 DSLSR 34.60 0.9289 30.59 0.8466 29.24 0.8085 28.73 0.8629 34.26 0.9486 Bicubic ×4 28.42 0.8104 26.00 0.7027 25.96 0.6675 23.14 0.6577 24.89 0.7866 SRCNN 30.48 0.8626 27.50 0.7513 26.90 0.7101 24.52 0.7221 27.58 0.8555 VDSR 31.35 0.8838 28.01 0.7674 27.29 0.7251 25.18 0.7524 28.83 0.8870 DRCN 31.53 0.8854 28.02 0.7670 27.23 0.7233 25.14 0.7510 28.98 0.8816 CARN 32.13 0.8937 28.60 0.7806 27.58 0.7349 26.07 0.7837 30.47 0.9084 IMDN 32.21 0.8948 28.58 0.7811 27.56 0.7353 26.04 0.7838 30.45 0.9075 EDSR 32.09 0.8938 28.58 0.7813 27.57 0.7357 26.04 0.7849 30.35 0.9067 SMSR 32.12 0.8932 28.55 0.7808 27.55 0.7351 26.11 0.7868 30.54 0.9085 RFDN 32.24 0.8952 28.61 0.7819 27.57 0.7360 26.11 0.7858 30.58 0.9089 SAFMN 32.18 0.8948 28.60 0.7813 27.58 0.7359 25.97 0.7809 30.43 0.9063 ESRT 32.19 0.8947 28.69 0.7833 27.69 0.7379 26.39 0.7962 30.75 0.9100 HNCT 32.31 0.8957 28.71 0.7834 27.63 0.7381 26.20 0.7896 30.70 0.9112 LBNet 32.29 0.8960 28.68 0.7832 27.62 0.7382 26.27 0.7906 30.76 0.9111 NGswin 32.33 0.8963 28.78 0.7859 27.66 0.7396 26.45 0.7963 30.80 0.9128 SwinIR-Light 32.44 0.8976 28.77 0.7858 27.69 0.7406 26.47 0.7980 30.92 0.9151 DSLSR 32.42 0.8976 28.81 0.7861 27.71 0.7406 26.63 0.8010 31.14 0.9162
Table 2. Evaluation metrics for various lightweight models at ×4 scale
View tableTable 2. Evaluation metrics for various lightweight models at ×4 scale
Method Parameter quantity /103 FLOPs /109 Set14 BSD100 Urban100 Manga109 PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM DRCN 1774 9788.0 28.02 0.7670 27.23 0.7233 25.14 0.7510 28.98 0.8816 VDSR 666 612.6 28.01 0.7674 27.29 0.7251 25.18 0.7524 28.83 0.8870 CARN 1592 90.9 28.60 0.7806 27.58 0.7349 26.07 0.7837 30.47 0.9084 SMSR 1006 41.6 28.55 0.7808 27.55 0.7351 26.11 0.7868 30.54 0.9085 EDSR 1518 114.0 28.58 0.7813 27.57 0.7357 26.04 0.7849 30.35 0.9067 NGswin 1019 36.4 28.78 0.7859 27.66 0.7396 26.45 0.7963 30.80 0.9128 SwinIR-Light 930 49.6 28.77 0.7858 27.69 0.7406 26.47 0.7980 30.92 0.9151 DSLSR 858 39.6 28.81 0.7861 27.71 0.7406 26.63 0.8010 31.14 0.9162
Table 3. Impact of different components on model performance
View tableTable 3. Impact of different components on model performance
Method Parameter quantity /103 Set5 Set14 BSD100 Urban100 Manga109 PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM Model1 1204 32.30 0.8965 28.65 0.7828 27.61 0.7370 26.15 0.7887 30.64 0.9107 Model2 258 32.07 0.8934 28.55 0.7798 27.52 0.7335 25.90 0.7778 30.25 0.9051 Model3 854 32.42 0.8980 28.78 0.7858 27.70 0.7406 26.51 0.7979 31.10 0.9160 Model4 858 32.42 0.8976 28.81 0.7861 27.71 0.7406 26.63 0.8010 31.14 0.9162
Table 4. Impact of different components in LFDB on model performance
View tableTable 4. Impact of different components in LFDB on model performance
3×3Conv 1×1Conv Shift8 DSEM Parameter quantity /103 PSNR /dB SSIM √ × × × 149.0 28.61 0.7819 × √ × × 29.1 27.52 0.7522 × √ √ × 29.1 28.55 0.7798 × √ √ √ 104.2 28.78 0.7858
Table 5. Impact of DSFM on DSEM performance
View tableTable 5. Impact of DSFM on DSEM performance
Method Set5 Set14 BSD100 Urban100 Manga109 PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM PSNR /dB SSIM DSLSR 32.42 0.8976 28.81 0.7861 27.71 0.7406 26.63 0.8010 31.14 0.9162 DSLSR without DSFM 32.44 0.8974 28.74 0.7851 27.65 0.7398 26.36 0.7940 30.78 0.9130
Tools
Get Citation
Copy Citation Text
Zhilin Gao, Jintao Wang, Qixiang Meng, Fanliang Bu. Lightweight Image Super-Resolution Reconstruction Incorporating Dual-Stream Feature Enhancement[J]. Laser & Optoelectronics Progress, 2025, 62(16): 1637002
Paper Information
Category: Digital Image Processing
Received: Jan. 3, 2025
Accepted: Mar. 11, 2025
Published Online: Aug. 6, 2025
The Author Email: Fanliang Bu (20051257@ppsuc.edu.cn)
CSTR:32186.14.LOP250444
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20