ISCSLP 2022智能座舱语音识别挑战赛数据集
收藏资源简介:
ISCSLP 2022智能座舱语音识别挑战赛数据集由西北工业大学计算机学院音频、语音和语言处理组创建,包含20小时的新能源汽车内语音数据,覆盖多种座舱声学条件和语言内容。数据集分为10小时的评估集和11小时的测试集,涉及空调控制、电话呼叫、音乐播放、导航等多种命令类型。创建过程中特别考虑了车内复杂的声学环境,旨在为车辆嵌入式及云端自动语音识别系统提供高质量数据,以提升驾驶安全和体验。
The ISCSLP 2022 Intelligent Cockpit Speech Recognition Challenge Dataset was developed by the Audio, Speech and Language Processing Group at the School of Computer Science, Northwestern Polytechnical University. It contains 20 hours of speech data collected inside new energy vehicles, covering various cockpit acoustic conditions and linguistic contents. The dataset is split into a 10-hour evaluation set and an 11-hour test set, which involve multiple command types such as air conditioning control, phone call, music playback, navigation and others. Special considerations were given to the complex in-vehicle acoustic environment during the dataset creation. It aims to provide high-quality data for in-vehicle embedded and cloud-based automatic speech recognition systems, so as to improve driving safety and user experience.

- 1The ISCSLP 2022 Intelligent Cockpit Speech Recognition Challenge (ICSRC): Dataset, Tracks, Baseline and Results西北工业大学计算机学院音频、语音和语言处理组(ASLP@NPU) · 2022年



