RescueSpeech
收藏资源简介:
RescueSpeech是一个专为搜索与救援(SAR)领域设计的德语语音数据集,由德国人工智能研究中心创建。该数据集包含约2小时的标注语音材料,来源于模拟救援演习中的真实语音记录。数据集的创建过程包括录音、降采样、分割和手动转录。RescueSpeech主要用于支持决策者和过程监控者在灾难情况下的工作,通过自动语音识别(ASR)输出与自然语言理解(NLU)组件结合,提取任务相关信息,辅助救援行动。
RescueSpeech is a German-language speech dataset tailored for the search and rescue (SAR) domain, developed by the German Research Center for Artificial Intelligence. It contains approximately 2 hours of annotated speech data sourced from real recordings made during simulated rescue exercises. The dataset was constructed through audio recording, downsampling, segmentation, and manual transcription. Primarily intended to support decision-makers and process monitors in their work during disaster scenarios, RescueSpeech combines automatic speech recognition (ASR) outputs with natural language understanding (NLU) components to extract task-relevant information and assist rescue operations.

- 1RescueSpeech: A German Corpus for Speech Recognition in Search and Rescue Domain德国人工智能研究中心 · 2023年



