改进深度卷积生成式对抗网络的文本生成图像

李云红; 朱绵云; 任劼; 苏雪平; 周小计; 于惠康

doi:10.13700/j.bh.1001-5965.2021.0588

改进深度卷积生成式对抗网络的文本生成图像

doi: 10.13700/j.bh.1001-5965.2021.0588

西安工程大学电子信息学院，西安 710048

基金项目: 国家自然科学基金(61902301)；陕西省自然科学基础研究计划重点项目(2022JZ-35)

详细信息

通讯作者:
E-mail：hitliyunhong@163.com

中图分类号: TP391.41
计量
- 文章访问数: 435
- HTML全文浏览量: 118
- PDF下载量: 103
- 被引次数: 0
出版历程
- 收稿日期: 2021-10-01
- 录用日期: 2021-12-24
- 网络出版日期: 2022-02-07
- 整期出版日期: 2023-08-31

Text-to-image synthesis based on modified deep convolutional generative adversarial network

School of Electronics and Information，Xi’an Polytechnic University，Xi’an 710048，China

Funds: National Natural Science Foundation of China (61902301); Key Project of Natural Science Basic Research Plan in Shaanxi Province of China (2022JZ-35)

More Information

Corresponding author: E-mail：hitliyunhong@163.com

摘要

摘要:
针对深度卷积生成式对抗网络(DCGAN) 模型高维文本输入表示的稀疏性导致以文本为条件生成的图像结构缺失和图像不真实的问题，提出了一种改进深度卷积生成式对抗网络模型CA-DCGAN。采用深度卷积网络和循环文本编码器对输入的文本进行编码，得到文本的特征向量表示。引入条件增强（CA）模型，通过文本特征向量的均值和协方差矩阵产生附加的条件变量，代替原来的高维文本特征向量。将条件变量与随机噪声结合作为生成器的输入，并在生成器的损失中额外加入KL损失正则化项，避免模型训练过拟合，使模型可以更好的收敛，在判别器中使用谱约束（SN）层，防止其梯度下降太快造成生成器与判别器不平衡训练而发生模式崩溃的问题。实验验证结果表明：所提模型在Oxford-102-flowers和CUB-200数据集上生成的图像质量较alignDRAW、GAN-CLS、GAN-INT-CLS、StackGAN(64×64)、StackGAN-v1(64×64)模型更好且接近于真实样本，初始得分值最低分别提高了10.9%和5.6%，最高分别提高了41.4%和37.5%，FID值最低分别降低了11.4%和8.4%，最高分别降低了43.9%和42.5%，进一步表明了所提模型的有效性。
- 深度卷积生成式对抗网络 /
- 文本生成图像 /
- 文本特征表示 /
- 条件增强 /
- KL正则化
Abstract:
When high-dimensional texts are adopted as input,images generated by the previously proposed deep convolutional generative adversarial network (DCGAN) model usually suffer from distortions and structure degradation due to the sparsity of texts, which seriously poses a negative impact on the generative performance. To address this issue, an improved deep convolutional generative adversarial network model, CA-DCGAN is proposed. Technically, a deep convolutional network and a recurrent text encoder are simultaneously employed to encode the input text so that the corresponding text embedding representation can be obtained. Then, a conditional augmentation (CA) model is introduced to generate an additional condition variable to replace the original high-dimensional text feature. Finally, the conditional variable and random noise are combined as the input of the generator. Meanwhile, to avoid over-fitting and promote the convergence,we introduce a KL regularization term into the generator’s loss. Moreover, we adopt a spectral normalization (SN) layer in the discriminator to prevent the mode collapse caused by the unbalanced training due to the fast gradient descent of the discriminator. The experimental verification results show that the proposed model on the Oxford-102-flowers and CUB-200 datasets is better than that of alignDRAW, GAN-CLS, GAN-INT-CLS, StackGAN (64×64), StackGAN-vl (64×64) in terms of the quality of generated images. The results show that the lowest inception score increased by 10.9% and 5.6% respectively, the highest inception score increased by 41.4% and 37.5% respectively, while the lowest FID index value decreased by 11.4% and 8.4% respectively,the highest FID index value decreased by 43.9% and 42.5% respectively,which further validate the effectiveness of the proposed method.
- deep convolutional generative adversarial network /
- text-to-image synthesis /
- text feature representation /
- conditional augmentation /
- KL regularization

HTML全文

图 1 GAN网络结构

Figure 1. GAN network structure

下载: 全尺寸图片幻灯片

图 2 CA-DCGAN宏观结构

Figure 2. Macroscopic structure of CA-DCGAN

下载: 全尺寸图片幻灯片

图 3 CA-DCGAN生成器网络结构

Figure 3. Generator network structure of CA-DCGAN

下载: 全尺寸图片幻灯片

图 4 CA-DCGAN判别器网络结构

Figure 4. Discriminator network structure of CA-DCGAN

下载: 全尺寸图片幻灯片

图 5 alignDRAW、GAN-CLS、GAN-INT-CLS、CA-DCGAN在Oxford-102-flowers数据集上生成结果比较

Figure 5. Comparison of generated results by alignDRAW, GAN-CLS,GAN-INT-CLS,CA-DCGAN models on Oxford-102-flowers dataset

下载: 全尺寸图片幻灯片

图 6 alignDRAW、GAN-CLS、GAN-INT-CLS、CA-DCGAN在CUB-200数据集上生成结果比较

Figure 6. Comparison of generated results by alignDRAW,GAN-CLS,GAN-INT-CLS,CA-DCGAN models on CUB-200 dataset

下载: 全尺寸图片幻灯片

图 7 StackGAN(64×64)、StackGAN-v1(64×64)与CA-DCGAN模型在Oxford-102-flowers和CUB-200数据集上生成结果比较

Figure 7. Comparison of generated results by StackGAN(64×64), StackGAN-v1(64×64) and CA-DCGAN models on Oxford-102-flowers dataset and CUB-200 dataset

下载: 全尺寸图片幻灯片

图 8 GAN-CLS、GAN-INT-CLS、CA-DCGAN在Oxford-102-flowers数据集上生成图像多样性对比

Figure 8. Comparison of image diversity generated by GAN-CLS,GAN-INT-CLS and CA-DCGAN on Oxford-102-flowers dataset

下载: 全尺寸图片幻灯片

图 9 GAN-CLS、GAN-INT-CLS、CA-DCGAN在CUB-200数据集上生成图像多样性对比

Figure 9. Comparison of image diversity generated by GAN-CLS,GAN-INT-CLS and CA-DCGAN on CUB-200 dataset

下载: 全尺寸图片幻灯片

表 1 实验数据集

Table 1. Experimental datasets

数据集		图像数量	文本描述/图像	类别
Oxford-102-flowers	训练集	7034	10	102
Oxford-102-flowers	测试集	1155	10	102
CUB-200	训练集	8855	10	200
CUB-200	测试集	2933	10	200

下载: 导出CSV

表 2 不同模型在Oxford-102-flowers和CUB-200数据集上G_IS值比较

Table 2. G_IS comparison of different models on Oxford-102-flowers and CUB-200 datasets

模型	Oxford-102-flowers	CUB-200
alignDRAW	2.15	2.32
GAN-CLS	2.49	2.64
GAN-INT-CLS	2.66	2.88
StackGAN^[20](64×64)	2.70	2.93
StackGAN-v1^[21](64×64)	2.74	3.02
CA-DCGAN	3.04	3.19

下载: 导出CSV

表 3 不同模型在Oxford-102-flowers和CUB-200数据集上G_FID值比较

Table 3. G_FID comparison of different models on Oxford-102-flowers and CUB-200 datasets

模型	Oxford-102	CUB-200
alignDRAW	96.87	88.54
GAN-CLS	88.78	79.25
GAN-INT-CLS	79.95	68.79
StackGAN^[20](64×64)	61.32	55.58
StackGAN-v1^[21](64×64)	43.02	35.11
CA-DCGAN	54.36	50.92

下载: 导出CSV

表 4 KL引入后指标变化对比

Table 4. Comparison of index change on KL’s introduction

模型	G_IS		G_FID
模型	Oxford-102-flowers	CUB-200	Oxford-102-flowers	CUB-200
Baseline	2.91	3.12	69.75	60.24
Baseline +KL，$ \lambda $=1	2.95	3.14	63.34	54.56
Baseline +KL，$ \lambda $=2	3.04	3.19	54.36	50.92
Baseline +KL，$ \lambda $=5	2.92	3.11	65.49	59.51
Baseline +KL，$ \lambda $=10	2.89	3.06	68.47	63.89

下载: 导出CSV

参考文献(21)

[1]	ZHOU K Y, YANG Y X, HOSPEDALES T, et al. Deep domain-adversarial image generation for domain generalisation[C]//34th AAAI Conference on Artificial Intelligence/32nd Innovative Applications of Artificial Intelligence Conference/10th AAAI Symposium on Educational Advances in Artificial Intelligence. Palo Alto: AAAI, 2020, 34: 13025-13032.
[2]	陆婷婷, 李潇, 张尧, 等. 基于三维点云模型的空间目标光学图像生成技术[J]. 北京航空航天大学学报, 2020, 46(2): 274-286. LU T T, LI X, ZHANG Y, et al. A technology for generation of space object optical image based on 3D point cloud model[J]. Journal of Beijing University of Aeronautics and Astronautics, 2020, 46(2): 274-286(in Chinese).
[3]	ZHANG Z, XIE Y, YANG L. Photographic text-to-image synthesis with a hierarchically-nested adversarial network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2018: 6199-6208.
[4]	牛蒙蒙, 沈明瑞, 秦波, 等. 基于GAN的刀具状态监测数据集增强方法[J]. 组合机床与自动化加工技术, 2021(4): 113-115. doi: 10.13462/j.cnki.mmtamt.2021.04.027 NIU M M, SHEN M R, QIN B, et al. A data augmentation method based on GAN in tool condition monitoring[J]. Combined Machine Tool and Automatic Machining Technology, 2021(4): 113-115(in Chinese). doi: 10.13462/j.cnki.mmtamt.2021.04.027
[5]	VENDROV I, KIROS R, FIDLER S, et al. Order-embeddings of images and language[EB/OL]. (2016-03-01)[2021-09-01]. https://arxiv.org/abs/1511.06361v3.
[6]	MANSIMOV E, PARISOTTO E, BA J L, et al. Generating images from captions with attention[EB/OL]. (2016-02-29)[2021-09-01]. https://arxiv.org/abs/1511.02793v2.
[7]	GREGOR K, DANIHELKA I, GRAVES A, et al. DRAW: A recurrent neural network for image generation[C]//Proceedings of the 32nd International Conference on Machine Learning. New York: ACM, 2015: 1462-1471.
[8]	REED S, VAN DEN OORD A, KALCHBRENNER N, et al. Generating interpretable images with controllable structure[C]//5th International Conference on Learning Representations, Appleton, WI: ICLR, 2016.
[9]	NGUYEN A, CLUNE J, BENGIO Y, et al. Plug & play generative networks: Conditional iterative generation of images in latent space[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2017: 17355648.
[10]	GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 2014: 2672-2680.
[11]	MIRZA M, OSINDERO S. Conditional generative adversarial nets[EB/OL]. (2014-10-06)[2021-09-01]. https://arxiv.org/abs/1411.1784.
[12]	SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 2002, 45(11): 2673-2681.
[13]	RADFORD A, METZ L, CHINTALA S. Unsupervised representation learning with deep convolutional generative adversarial networks[C]//4th International Conference on Learning Representations, Appleton, WI: ICLR, 2016.
[14]	REED S, AKATA Z, YAN X, et al. Generative adversarial text to image synthesis[C]//Proceedings of the 33rd International Conference on Machine Learning. New York: ACM, 2016: 1060-1069.
[15]	REED S, AKATA Z, LEE H, et al. Learning deep representations of fine-grained visual descriptions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2016: 49-58.
[16]	NILSBACK M E, ZISSERMAN A. Automated flower classification over a large number of classes[C]//Proceedings of the IEEE Conference on Computer Vision, Graphics and Image Processing. Piscataway: IEEE Press, 2008: 722-729.
[17]	WAH C, BRANSON S, WELINDER P, et al. The Caltech-UCSD birds-200-2011 dataset: CNS-TR-2011-001[R]. Pasadena: California Institute of Technology, 2011.
[18]	SALIMANS T, GOODFELLOW I, ZAREMBA W, et al. Improved techniques for training GANs[C]//30th Conference on Neural Information Processing Systems, Cambridge: MIT Press, 2016: 2234-2242.
[19]	HEUSEL M, RAMSAUER H, UNTERTHINER T, et al. GANs trained by a two time-scale update rule converge to a local Nash equilibrium[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 2017: 6626-6637.
[20]	ZHANG H, XU T, LI H, et al. StackGAN: Text to photo-realistic image synthesis with stacked generative adversarial networks[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway: IEEE Press, 2017: 5907-5915.
[21]	ZHANG H, XU T, LI H, et al. StackGAN++: Realistic image synthesis with stacked generative adversarial networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(8): 1947-1962. doi: 10.1109/TPAMI.2018.2856256