研究人员正在开发先进的音频和视听深度伪造检测方法，以应对部分操纵的语音和歌声等挑战。一种名为CAFNet的新架构能够联合分类音频为真实、完全伪造或半真半假，并高精度地定位被操纵的片段。另一个框架T-AVFD利用文本指导，通过学习可泛化的真实性模式，提高了在说话和歌唱场景下的检测能力。此外，一项大规模研究表明，尽管人类检测音频深度伪造的能力没有显著提高，但人们对真实语音的信任度却在下降，这表明音频真实性的信任正在广泛受到侵蚀。 AI
影响 先进的检测方法和对人类感知的研究对于维护音频信任和打击生成式AI的恶意使用至关重要。
排序理由 该集群包含三篇学术论文，详细介绍了音频和视听深度伪造检测技术的新研究及其对人类感知的影响。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.29531v1 Announce Type: cross Abstract: Audio deepfake detection is well-studied as a binary problem, but partially manipulated speech, where a short synthesised segment is spliced into an otherwise genuine utterance, poses a harder and more realistic threat. Detecting …
arXiv:2605.27944v1 Announce Type: new Abstract: With rapid advances in audio-visual generative models, reliable forgery detection becomes increasingly critical. Existing methods for audio-visual deepfake detection typically rely on cross-modal inconsistencies. In singing, rhythmi…
arXiv:2605.26136v1 Announce Type: cross Abstract: Audio deepfakes have improved rapidly recently, yet their effect on human trust in real speech remains unstudied. We present the largest listening study on audio deepfake perception to date, collecting 35,532 judgments from 1,768 …