arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2302.08841cs.SDcs.CVcs.LGcs.MMeess.AS
收藏

基于多任务学习的唇到语音合成

Lip-to-Speech Synthesis in the Wild with Multi-task Learning

KAIST(韩国科学技术院)

Minsu Kim, Joanna Hong, Yong Man Ro

摘要:

最近的研究表明,在唇到语音合成,旨在重建语音从视觉信息单独令人印象深刻的性能。然而,由于指导模型推断正确内容的监督不足,他们一直无法在野外合成准确的语音。与以前的方法不同,在本文中,我们开发了一个功能强大的Lip2Speech方法,即使在野外环境中,也可以从输入的嘴唇运动中重建具有正确内容的语音。为此,我们设计了多任务学习,使用多模态监督来指导模型,即,文本和音频,以补充声学特征重建损失的单词表示的不足。因此,所提出的框架带来了合成语音的优势,包含多个扬声器的权利内容与不受约束的句子。我们使用LRS2,LRS3和LRW数据集验证了所提出的方法的有效性。

英文摘要:

Recent studies have shown impressive performance in Lip-to-speech synthesis that aims to reconstruct speech from visual information alone. However, they have been suffering from synthesizing accurate speech in the wild, due to insufficient supervision for guiding the model to infer the correct content. Distinct from the previous methods, in this paper, we develop a powerful Lip2Speech method that can reconstruct speech with correct contents from the input lip movements, even in a wild environment. To this end, we design multi-task learning that guides the model using multimodal supervision, i.e., text and audio, to complement the insufficient word representations of acoustic feature reconstruction loss. Thus, the proposed framework brings the advantage of synthesizing speech containing the right content of multiple speakers with unconstrained sentences. We verify the effectiveness of the proposed method using LRS2, LRS3, and LRW datasets.

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。

↑