利用增强的类儿童语音改进儿童语音识别
Improving child speech recognition with augmented child-like speech
AI总结：
针对儿童语音数据稀缺问题，本文提出儿童到儿童语音转换增强方法，发现跨语言VC可显著降低ASR词错误率，并确定不同训练方式下的最佳增强倍数。
摘要：
最先进的自动语音识别（ASR）系统在儿童语音上表现欠佳。儿童语音数据的稀缺限制了儿童语音识别（CSR）的发展。因此，我们分别研究了利用数据集中现有儿童说话人的儿童到儿童语音转换（VC），以及通过单语和跨语言（荷兰语到德语）VC利用额外（新）儿童说话人的方法。结果表明，跨语言儿童到儿童VC显著提升了儿童ASR性能。关于儿童到儿童跨语言VC生成数据数量对微调（FT）ASR模型影响的实验显示：FT-Conformer模型和FT-Whisper模型在两倍增强时取得最佳结果，与基线相比词错误率（WER）绝对降低约3%；而从头训练的模型在六倍增强时效果最佳，WER绝对改善3.6%。此外，使用少量“高质量”VC生成数据即可达到与最佳FT模型相似的结果。
英文摘要：
State-of-the-art ASRs show suboptimal performance for child speech. The scarcity of child speech limits the development of child speech recognition (CSR). Therefore, we studied child-to-child voice conversion (VC) from existing child speakers in the dataset and additional (new) child speakers via monolingual and cross-lingual (Dutch-to-German) VC, respectively. The results showed that cross-lingual child-to-child VC significantly improved child ASR performance. Experiments on the impact of the quantity of child-to-child cross-lingual VC-generated data on fine-tuning (FT) ASR models gave the best results with two-fold augmentation for our FT-Conformer model and FT-Whisper model which reduced WERs with ~3% absolute compared to the baseline, and with six-fold augmentation for the model trained from scratch, which improved by an absolute 3.6% WER. Moreover, using a small amount of "high-quality" VC-generated data achieved similar results to those of our best-FT models.
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。
