arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.11437cs.LGstat.ML
收藏

用于文本到图像合成的高效神经结构

Efficient Neural Architecture for Text-to-Image Synthesis

Douglas M. Souza, Jônatas Wehrmann, Duncan D. Ruiz

摘要:

文本到图像合成是从文本描述生成图像的任务。图像生成本身就是一项具有挑战性的任务。当我们结合图像生成和文本时,我们将复杂性提升到一个新的水平:我们需要结合来自两种不同模态的数据。最近在文本到图像合成方面的大多数工作在涉及神经架构时都遵循类似的方法。由于上述困难,加上在高分辨率下训练GANs的固有困难,大多数方法都采用了多阶段训练策略。在本文中,我们改变了目前在文本到图像方法中使用的架构范式,并证明了一个有效的神经架构可以使用单个生成器和单个训练器的单阶段训练来实现最先进的性能。我们通过应用深度残差网络以及一种新的句子插值策略来实现这一点,该策略可以学习一个平滑的条件空间。最后,我们的工作为文本到图像的研究指出了一个新的方向,最近还没有尝试新的神经架构。

英文摘要:

Text-to-image synthesis is the task of generating images from text descriptions. Image generation, by itself, is a challenging task. When we combine image generation and text, we bring complexity to a new level: we need to combine data from two different modalities. Most of recent works in text-to-image synthesis follow a similar approach when it comes to neural architectures. Due to aforementioned difficulties, plus the inherent difficulty of training GANs at high resolutions, most methods have adopted a multi-stage training strategy. In this paper we shift the architectural paradigm currently used in text-to-image methods and show that an effective neural architecture can achieve state-of-the-art performance using a single stage training with a single generator and a single discriminator. We do so by applying deep residual networks along with a novel sentence interpolation strategy that enables learning a smooth conditional space. Finally, our work points a new direction for text-to-image research, which has not experimented with novel neural architectures recently.

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。

↑