Files
icassp2027-experiment/literature_notes/41_visual_turn_taking_noise.md
2026-08-30 15:36:28 +08:00

14 lines
1.0 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# Visual Cues Support Robust Turn-taking Prediction in Noise
- 作者:Sam O’Connor Russell、Naomi Harte
- 发表:Interspeech 2025
- 阅读深度:重点输入、噪声实验及局限章节
- 记录日期:2026-08-28
- 来源:[ISCA](https://www.isca-archive.org/interspeech_2025/oconnorrussell25_interspeech.html)、[论文](https://www.isca-archive.org/interspeech_2025/oconnorrussell25_interspeech.pdf)
## 摘要与项目关系
研究预测式轮次模型对噪声的脆弱性,显示噪声训练能帮助多模态模型利用视觉,但对未知噪声和不准转写的泛化有限。视觉特征包括脸部动作、gaze、头姿等;数据为双人视频会议和逐人音轨。
我们可增加自然重叠、低信噪比、人物可见性分层,避免只看总体 AP;噪声增强应独立于正常社交互动主实验。其双路音频与对齐标注条件不等于 EasyCom 当前的单路混合输入。人为错配视觉导致下降也可能只是分布外错误,不能单独证明社会关系理解。