用语音检测分段并行转录，实现法庭命令式口述的高效离线语音识别
Scalable Offline ASR for Command-Style Dictation in Courtrooms
- 通过声学活动检测分割音频，用Whisper模型并行处理提升效率
- 在印度约15%的法庭部署，用户并发增加时延迟持续降低
- 开源兼容多种语音识别模型，适合司法等高并发低延迟场景
我们提出一个开源框架，用于解决命令式口述中资源密集型在线系统与高延迟批处理之间的差距。该方法利用语音活动检测（VAD）对音频进行分段，并使用Whisper模型并行转录这些片段，实现跨音频的高效多路复用。与SuperWhisper等专有系统不同，本框架还兼容大多数ASR架构，包括广泛使用的基于CTC的模型。多路复用技术在实际环境中最大化计算资源利用率，已在约15%的印度法庭部署。在实时数据上的评估显示，随着用户并发量增加，延迟持续下降，优于串行批处理。现场演示将展示开源实现，并允许观众实时交互。
原文摘要 · Abstract (English)
We propose an open-source framework for Command-style dictation that addresses the gap between resource-intensive Online systems and high-latency Batch processing. Our approach uses Voice Activity Detection (VAD) to segment audio and transcribes these segments in parallel using Whisper models, enabling efficient multiplexing across audios. Unlike proprietary systems like SuperWhisper, this framework is also compatible with most ASR architectures, including widely used CTC-based models. Our multiplexing technique maximizes compute utilization in real-world settings, as demonstrated by its deployment in around 15% of India's courtrooms. Evaluations on live data show consistent latency reduction as user concurrency increases, compared to sequential batch processing. The live demonstration will showcase our open-sourced implementation and allow attendees to interact with it in real-time.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。