arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2005.06968cs.LGcs.CLcs.CV
收藏

S2IGAN:通过对抗学习生成语音到图像

S2IGAN: Speech-to-Image Generation via Adversarial Learning

Xi'an Jiaotong University(西安交通大学);Delft University of Technology(代尔夫特理工大学);Zhejiang University(浙江大学)

Xinsheng Wang, Tingting Qiao, Jihua Zhu, Alan Hanjalic, Odette Scharenborg

摘要:

据估计,世界上一半的语言没有书面形式,这使得这些语言不可能从任何现有的基于文本的技术中受益。在本文中,语音到图像生成(S2 IG)框架,提出了将语音描述转换为照片般逼真的图像,而不使用任何文本信息,从而使非书面语言可能受益于这项技术。提出的S2 IG框架,命名为S2 IGAN,由一个语音嵌入网络(SEN)和一个关系监督密集堆叠生成模型(RDG)。SEN在相应视觉信息的监督下学习语音嵌入。在SEN产生的语音嵌入的条件下,建议的RDG合成的图像是语义上与相应的语音描述一致。在两个公共基准数据集CUB和Oxford-102上的实验证明了所提出的S2 IGAN在从语音信号合成高质量和语义一致的图像方面的有效性,为S2 IG任务提供了良好的性能和坚实的基线。

英文摘要:

An estimated half of the world's languages do not have a written form, making it impossible for these languages to benefit from any existing text-based technologies. In this paper, a speech-to-image generation (S2IG) framework is proposed which translates speech descriptions to photo-realistic images without using any text information, thus allowing unwritten languages to potentially benefit from this technology. The proposed S2IG framework, named S2IGAN, consists of a speech embedding network (SEN) and a relation-supervised densely-stacked generative model (RDG). SEN learns the speech embedding with the supervision of the corresponding visual information. Conditioned on the speech embedding produced by SEN, the proposed RDG synthesizes images that are semantically consistent with the corresponding speech descriptions. Extensive experiments on two public benchmark datasets CUB and Oxford-102 demonstrate the effectiveness of the proposed S2IGAN on synthesizing high-quality and semantically-consistent images from the speech signal, yielding a good performance and a solid baseline for the S2IG task.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑