Journal of Computer Applications ›› 2025, Vol. 45 ›› Issue (10): 3067-3073.DOI: 10.11772/j.issn.1001-9081.2024101535
• Artificial intelligence •
Guoyu XU, Xiaolong YAN(), Yidan ZHANG
Received:
2024-10-31
Revised:
2024-12-12
Accepted:
2024-12-20
Online:
2025-03-18
Published:
2025-10-10
Contact:
Xiaolong YAN
About author:
XU Guoyu, born in 1982, Ph. D., associate professor. His research interests include deep learning.Supported by:
通讯作者:
闫晓龙
作者简介:
徐国愚(1982—),男,安徽庐江人,副教授,博士,CCF高级会员,主要研究方向:深度学习基金资助:
CLC Number:
Guoyu XU, Xiaolong YAN, Yidan ZHANG. DU-FastGAN: lightweight generative adversarial network based on dynamic-upsample[J]. Journal of Computer Applications, 2025, 45(10): 3067-3073.
徐国愚, 闫晓龙, 张一丹. 基于动态上采样的轻量级生成对抗网络DU-FastGAN[J]. 《计算机应用》唯一官方网站, 2025, 45(10): 3067-3073.
Add to citation manager EndNote|Ris|BibTeX
URL: https://www.joca.cn/EN/10.11772/j.issn.1001-9081.2024101535
分辨率 | 子数据集 | 样本数 |
---|---|---|
256×256 | Dog | 389 |
Cat | 160 | |
Human Face | 100 | |
Panda | 101 | |
1 024×1 024 | Pokemon | 833 |
Art-Painting | 124 | |
Fauvism | 134 | |
Moongate | 137 | |
Shells | 97 | |
Skulls | 64 |
Tab. 1 Details of datasets
分辨率 | 子数据集 | 样本数 |
---|---|---|
256×256 | Dog | 389 |
Cat | 160 | |
Human Face | 100 | |
Panda | 101 | |
1 024×1 024 | Pokemon | 833 |
Art-Painting | 124 | |
Fauvism | 134 | |
Moongate | 137 | |
Shells | 97 | |
Skulls | 64 |
模型 | Epoch/104 | Dog | Cat | Human Face | Panda | ||||
---|---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | ||
FastGAN | 6 | 101.81 | 0.635 5 | 129.74 | 0.605 4 | 50.55 | 0.532 7 | 11.32 | 0.522 2 |
MixDL | 12 | 155.61 | 0.603 3 | 147.93 | 0.587 0 | 61.57 | 0.503 7 | 13.99 | 0.479 6 |
RCL-master | 20 | 99.26 | 0.637 1 | 131.94 | 0.610 3 | 51.48 | 0.561 3 | 11.07 | 0.528 3 |
DU-FastGAN | 6 | 92.29 | 0.647 1 | 112.17 | 0.617 8 | 48.71 | 0.594 5 | 10.40 | 0.540 5 |
Tab. 2 Comparison of results on small sample datasets at 256 × 256 resolution
模型 | Epoch/104 | Dog | Cat | Human Face | Panda | ||||
---|---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | ||
FastGAN | 6 | 101.81 | 0.635 5 | 129.74 | 0.605 4 | 50.55 | 0.532 7 | 11.32 | 0.522 2 |
MixDL | 12 | 155.61 | 0.603 3 | 147.93 | 0.587 0 | 61.57 | 0.503 7 | 13.99 | 0.479 6 |
RCL-master | 20 | 99.26 | 0.637 1 | 131.94 | 0.610 3 | 51.48 | 0.561 3 | 11.07 | 0.528 3 |
DU-FastGAN | 6 | 92.29 | 0.647 1 | 112.17 | 0.617 8 | 48.71 | 0.594 5 | 10.40 | 0.540 5 |
模型 | Epoch/104 | Pokemon | Art-Painting | Fauvism | Moongate | Shells | Skulls | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | ||
FastGAN | 6 | 75.77 | 0.567 | 51.58 | 0.746 | 194.11 | 0.775 | 139.94 | 0.709 | 296.84 | 0.433 | 202.59 | 0.620 |
MixDL | 12 | 98.42 | 0.514 | 63.33 | 0.720 | 204.01 | 0.721 | 177.29 | 0.675 | 304.91 | 0.401 | 253.20 | 0.607 |
RCL-master | 20 | 70.13 | 0.633 | 49.67 | 0.732 | 189.16 | 0.741 | 140.84 | 0.697 | 229.27 | 0.458 | 190.57 | 0.640 |
DU-FastGAN | 6 | 75.69 | 0.607 | 51.36 | 0.744 | 185.33 | 0.800 | 134.87 | 0.712 | 227.17 | 0.464 | 160.69 | 0.667 |
Tab. 3 Comparison of results on small sample datasets at 1 024 × 1 024 resolution
模型 | Epoch/104 | Pokemon | Art-Painting | Fauvism | Moongate | Shells | Skulls | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | ||
FastGAN | 6 | 75.77 | 0.567 | 51.58 | 0.746 | 194.11 | 0.775 | 139.94 | 0.709 | 296.84 | 0.433 | 202.59 | 0.620 |
MixDL | 12 | 98.42 | 0.514 | 63.33 | 0.720 | 204.01 | 0.721 | 177.29 | 0.675 | 304.91 | 0.401 | 253.20 | 0.607 |
RCL-master | 20 | 70.13 | 0.633 | 49.67 | 0.732 | 189.16 | 0.741 | 140.84 | 0.697 | 229.27 | 0.458 | 190.57 | 0.640 |
DU-FastGAN | 6 | 75.69 | 0.607 | 51.36 | 0.744 | 185.33 | 0.800 | 134.87 | 0.712 | 227.17 | 0.464 | 160.69 | 0.667 |
模型 | Dog | Skulls | Human Face | |||
---|---|---|---|---|---|---|
训练 | 生成 | 训练 | 生成 | 训练 | 生成 | |
FastGAN | 498 | 0.40 | 503 | 0.22 | 505 | 0.12 |
MixDL | 727 | 0.30 | 723 | 0.30 | 719 | 0.27 |
RCL-master | 13 192 | 0.33 | 13 217 | 0.27 | 13 199 | 0.27 |
DU-FastGAN | 599 | 0.18 | 593 | 0.30 | 597 | 0.18 |
Tab. 4 Comparison of model training and sample generation time
模型 | Dog | Skulls | Human Face | |||
---|---|---|---|---|---|---|
训练 | 生成 | 训练 | 生成 | 训练 | 生成 | |
FastGAN | 498 | 0.40 | 503 | 0.22 | 505 | 0.12 |
MixDL | 727 | 0.30 | 723 | 0.30 | 719 | 0.27 |
RCL-master | 13 192 | 0.33 | 13 217 | 0.27 | 13 199 | 0.27 |
DU-FastGAN | 599 | 0.18 | 593 | 0.30 | 597 | 0.18 |
模型 | Dog | Skulls | Human Face | Panda | ||||
---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | |
Baseline | 184.11 | 0.635 5 | 203.44 | 0.619 9 | 50.55 | 0.532 7 | 11.32 | 0.522 2 |
F1 | 87.71 | 0.640 6 | 169.66 | 0.621 3 | 49.43 | 0.564 4 | 11.07 | 0.536 7 |
F2 | 113.76 | 0.635 5 | 181.03 | 0.632 7 | 48.92 | 0.573 2 | 10.96 | 0.529 1 |
F3 | 107.63 | 0.631 5 | 172.39 | 0.651 7 | 49.11 | 0.570 3 | 11.14 | 0.535 8 |
F4 | 83.69 | 0.641 2 | 163.37 | 0.638 4 | 49.27 | 0.578 2 | 10.90 | 0.537 3 |
F5 | 97.68 | 0.637 5 | 171.51 | 0.645 2 | 48.86 | 0.579 5 | 10.74 | 0.531 6 |
F6 | 90.04 | 0.645 1 | 162.94 | 0.659 3 | 48.91 | 0.586 7 | 10.97 | 0.538 1 |
DU-FastGAN | 82.64 | 0.647 1 | 160.69 | 0.666 6 | 48.71 | 0.594 5 | 10.40 | 0.540 5 |
Tab. 5 Ablation experimental results
模型 | Dog | Skulls | Human Face | Panda | ||||
---|---|---|---|---|---|---|---|---|
FID | LPIPS | FID | LPIPS | FID | LPIPS | FID | LPIPS | |
Baseline | 184.11 | 0.635 5 | 203.44 | 0.619 9 | 50.55 | 0.532 7 | 11.32 | 0.522 2 |
F1 | 87.71 | 0.640 6 | 169.66 | 0.621 3 | 49.43 | 0.564 4 | 11.07 | 0.536 7 |
F2 | 113.76 | 0.635 5 | 181.03 | 0.632 7 | 48.92 | 0.573 2 | 10.96 | 0.529 1 |
F3 | 107.63 | 0.631 5 | 172.39 | 0.651 7 | 49.11 | 0.570 3 | 11.14 | 0.535 8 |
F4 | 83.69 | 0.641 2 | 163.37 | 0.638 4 | 49.27 | 0.578 2 | 10.90 | 0.537 3 |
F5 | 97.68 | 0.637 5 | 171.51 | 0.645 2 | 48.86 | 0.579 5 | 10.74 | 0.531 6 |
F6 | 90.04 | 0.645 1 | 162.94 | 0.659 3 | 48.91 | 0.586 7 | 10.97 | 0.538 1 |
DU-FastGAN | 82.64 | 0.647 1 | 160.69 | 0.666 6 | 48.71 | 0.594 5 | 10.40 | 0.540 5 |
[1] | PORKODI S P, SARADA V, MAIK V, et al. Generic image application using GANs (Generative Adversarial Networks): a review[J]. Evolving Systems, 2023, 14(5): 903-917. |
[2] | 杨玮,钟名锋,杨根,等. 基于NVAE和OB-Mix的小样本数据增强方法[J]. 计算机工程与应用, 2024, 60(2):103-112. |
YANG W, ZHONG M F, YANG G, et al. Few samples data augmentation method based on NVAE and OB-Mix[J]. Computer Engineering and Applications, 2024, 60(2):103-112. | |
[3] | WEI J, WANG Q, ZHAO Z. Generative adversarial network based on Poincaré distance similarity constraint: focusing on overfitting problem caused by finite training data[J]. Applied Soft Computing, 2024, 151: No.111147. |
[4] | SNIDER E J, HERNANDEZ-TORRES S I, HENNESSEY R. Using ultrasound image augmentation and ensemble predictions to prevent machine-learning model overfitting[J]. Diagnostics, 2023, 13(3): No.417. |
[5] | NAVEED H, ANWAR S, HAYAT M, et al. Survey: image mixing and deleting for data augmentation[J]. Engineering Applications of Artificial Intelligence, 2024, 131: No.107791. |
[6] | 姜文涛,刘玉薇,张晟翀. 随机通道扰动的图像数据增强方法[J]. 计算机科学与探索, 2024, 18(11): 2980-2995. |
JIANG W T, LIU Y W, ZHANG S C. Image data augmentation method for random channel perturbation[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(11): 2980-2995. | |
[7] | AGGARWAL A, MITTAL M, BATTINENI G. Generative adversarial network: an overview of theory and applications[J]. International Journal of Information Management Data Insights, 2021, 1(1): No.100004. |
[8] | BISWAS A, NASIM M A AL, IMRAN A, et al. Generative adversarial networks for data augmentation[M]// ZHENG B, ANDREI S, SARKER M K, et al. Data driven approaches on medical imaging. Cham: Springer, 2023: 159-177. |
[9] | MIRZA M, OSINDERO S. Conditional generative adversarial nets[EB/OL]. [2024-07-03].. |
[10] | ARJOVSKY M, CHINTALA S, BOTTOU L. Wasserstein generative adversarial networks[C]// Proceedings of the 34th International Conference on Machine Learning. New York: JMLR.org, 2017: 214-223. |
[11] | DONAHUE J, SIMONYAN K. Large scale adversarial representation learning[C]// Proceedings of the 33rd International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2019: 10542-10552. |
[12] | KARRAS T, AILA T, LAINE S, et al. Progressive growing of GANs for improved quality, stability, and variation[EB/OL]. [2024-05-02].. |
[13] | BHATTAD A, McKEE D, HOIEM D, et al. StyleGAN knows normal, depth, albedo, and more[C]// Proceedings of the 37th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2023: 73082-73103. |
[14] | LIU B, ZHU Y, SONG K, et al. Towards faster and stabilized GAN training for high-fidelity few-shot image synthesis[EB/OL]. [2024-05-02].. |
[15] | KONG C, KIM J, HAN D, et al. Few-shot image generation with mixup-based distance learning[C]// Proceedings of the 2022 European Conference on Computer Vision, LNCS 13675. Cham: Springer, 2022: 563-580. |
[16] | GOU Y, LI M, ZHANG Y, et al. Few-shot image generation with reverse contrastive learning[J]. Neural Networks, 2024, 169: 154-164. |
[17] | HU H, LIU Z, LI L, et al. Pixel-wise smoothing for certified robustness against camera motion perturbations[C]// Proceedings of the 27th International Conference on Artificial Intelligence and Statistics. New York: JMLR.org, 2024: 217-225. |
[18] | ALLAHYANI M, ALSULAMI R, ALWAFI T, et al. DivGAN: a diversity enforcing generative adversarial network for mode collapse reduction[J]. Artificial Intelligence, 2023, 317: No.103863. |
[19] | DENG J, WANG S, YE J, et al. DGRM: diffusion-GAN recommendation model to alleviate the mode collapse problem in sparse environments[J]. Pattern Recognition, 2024, 155: No.110692. |
[20] | DESALEGN L, JIFARA W. HARA-GAN: hybrid attention and relative average discriminator based generative adversarial network for MR image reconstruction[J]. IEEE Access, 2024, 12: 23240-23251. |
[21] | ROY A, DASGUPTA D. DRD-GAN: a novel distributed conditional Wasserstein deep convolutional relativistic discriminator GAN with improved convergence[J]. ACM Transactions on Probabilistic Machine Learning, 2025, 1(1): No.6. |
[22] | YU Y, ZHANG W, DENG Y. Fréchet Inception Distance (FID) for evaluating GANs[EB/OL]. [2024-04-23].. |
[23] | GHAZANFARI S, GARG S, KRISHNAMURTHY P, et al. R-LPIPS: an adversarially robust perceptual similarity metric[EB/OL]. [2024-08-11].. |
[1] | Yilin DENG, Fajiang YU. Pseudo random number generator based on LSTM and separable self-attention mechanism [J]. Journal of Computer Applications, 2025, 45(9): 2893-2901. |
[2] | Yiming LIANG, Jing FAN, Wenze CHAI. Multi-scale feature fusion sentiment classification based on bidirectional cross attention [J]. Journal of Computer Applications, 2025, 45(9): 2773-2782. |
[3] | Chuang WANG, Lu YU, Jianwei CHEN, Cheng PAN, Wenbo DU. Review of open set domain adaptation [J]. Journal of Computer Applications, 2025, 45(9): 2727-2736. |
[4] | Jin ZHOU, Yuzhi LI, Xu ZHANG, Shuo GAO, Li ZHANG, Jiachuan SHENG. Modulation recognition network for complex electromagnetic environments [J]. Journal of Computer Applications, 2025, 45(8): 2672-2682. |
[5] | Chaoying JIANG, Qian LI, Ning LIU, Lei LIU, Lizhen CUI. Readmission prediction model based on graph contrastive learning [J]. Journal of Computer Applications, 2025, 45(6): 1784-1792. |
[6] | Daoquan LI, Zheng XU, Sihui CHEN, Jiayu LIU. Network traffic classification model integrating variational autoencoder and AdaBoost-CNN [J]. Journal of Computer Applications, 2025, 45(6): 1841-1848. |
[7] | Ying HUANG, Shengmei GAO, Guang CHEN, Su LIU. Low-light image enhancement network combining signal-to-noise ratio guided dual-branch structure and histogram equalization [J]. Journal of Computer Applications, 2025, 45(6): 1971-1979. |
[8] | Hui LI, Bingzhi JIA, Chenxi WANG, Ziyu DONG, Jilong LI, Zhaoman ZHONG, Yanyan CHEN. Generative adversarial network underwater image enhancement model based on Swin Transformer [J]. Journal of Computer Applications, 2025, 45(5): 1439-1446. |
[9] | Xueying LI, Kun YANG, Guoqing TU, Shubo LIU. Adversarial sample generation method for time-series data based on local augmentation [J]. Journal of Computer Applications, 2025, 45(5): 1573-1581. |
[10] | Renjie TIAN, Mingli JING, Long JIAO, Fei WANG. Recommendation algorithm of graph contrastive learning based on hybrid negative sampling [J]. Journal of Computer Applications, 2025, 45(4): 1053-1060. |
[11] | Haitao SUN, Jiayu LIN, Zuhong LIANG, Jie GUO. Data augmentation technique incorporating label confusion for Chinese text classification [J]. Journal of Computer Applications, 2025, 45(4): 1113-1119. |
[12] | Lihu PAN, Shouxin PENG, Rui ZHANG, Zhiyang XUE, Xuzhen MAO. Video anomaly detection for moving foreground regions [J]. Journal of Computer Applications, 2025, 45(4): 1300-1309. |
[13] | Chenwei SUN, Junli HOU, Xianggen LIU, Jiancheng LYU. Large language model prompt generation method for engineering drawing understanding [J]. Journal of Computer Applications, 2025, 45(3): 801-807. |
[14] | Hong SHANGGUAN, Huiying REN, Xiong ZHANG, Xinglong HAN, Zhiguo GUI, Yanling WANG. Low-dose CT denoising model based on dual encoder-decoder generative adversarial network [J]. Journal of Computer Applications, 2025, 45(2): 624-632. |
[15] | Kun FU, Shicong YING, Tingting ZHENG, Jiajie QU, Jingyuan CUI, Jianwei LI. Graph data augmentation method for few-shot node classification [J]. Journal of Computer Applications, 2025, 45(2): 392-402. |
Viewed | ||||||
Full text |
|
|||||
Abstract |
|
|||||