| 基于生成对抗文本的人脸图像翻译 |
 点此下载全文 |
| 引用本文：何佩林1，石跃祥1，2?覮，成洁1.基于生成对抗文本的人脸图像翻译[J].计算技术与自动化,2018,(4):77-82 |
| 摘要点击次数: 1508 |
| 全文下载次数: 0 |
|
|
| 中文摘要:为了使得人脸图像翻译方法具有更好的翻译效果，提出了一种基于生成对抗文本的人脸图像翻译方法（T-GAN）。根据人脸的特殊性，利用深度对称结构联合编码方法，提取人脸所对应的文本描述特征。然后结合生成对抗“博弈”的思想，强迫判别网络判断生成的图像是否符合文本描述，让判别网络不仅仅能够学习生成图像和输入图像之间的关系，也能够学习生成图像和文本描述的对应关系，达到强化训练的效果。实验结果表明，本方法具有使用价值，在各种肤色、发色人脸图像翻译任务都给出了高质量的图像翻译结果，并与其他图像翻译方法相比较，翻译效果更好。 |
| 中文关键词:人脸图像翻译 生成对抗文本 深度对称结构联合编码 |
| |
| Face Image Translation Based on Generative Adversarial Text |
|
|
| Abstract:In order to promote the translation effect of the?face image translation method， we propose a face image translation method based on generative adversarial text （T-GAN）. According to the particularity of human face， using the method of deep symmetric structured joint embedding， the corresponding text description features of face are extracted. Then combining with the idea of generative adversarial “game”， which forcing the discriminant network to judge whether the generated image accords with the text descriptions， so that the discriminant network not only learn the relationship between the generated image and the input image， but also learn the corresponding relation between the generated image and the text descriptions to achieve the effect of intensive training. Experimental results show that the proposed method is applicable with image translation results of high quality in all kinds of skin color and hair color face image translation tasks，and the translation effect is better that compared with other image translation methods. |
| keywords:face image translation generative adversarial text deep symmetric structured joint embedding |
| 查看全文 查看/发表评论 下载pdf阅读器 |
