DeepMine数据库

DeepMine数据库是一个专为波斯语和英语设计的语音数据库，旨在构建和评估文本依赖、文本提示和文本独立的发言人验证系统，以及波斯语语音识别系统。该数据库包含超过1850名发言人和54万条录音，总时长超过480小时，并已全部转录。它是第一个公开的大型波斯语发言人验证数据库，也是英语中最大的公开文本依赖和文本提示发言人验证数据库，以及最大的公开评估数据集用于文本独立发言人验证。数据库覆盖了年龄、性别和口音的广泛范围。DeepMine项目于2017年初启动，经过数据库设计、开发Android和服务器应用程序后，数据收集于2017年年中开始，最终于2018年底完成，并在2019年初发布了清理后的最终版本。该数据库主要用于文本依赖发言人验证，但也适用于波斯语的自动语音识别（ASR）模型训练。

The DeepMine Database is a speech database designed specifically for Persian and English, aiming to construct and evaluate text-dependent, text-prompted, and text-independent speaker verification systems, as well as Persian speech recognition systems. This database contains over 1,850 speakers and 540,000 audio recordings, with a total duration exceeding 480 hours, and all recordings have been fully transcribed. It is the first publicly available large-scale Persian speaker verification database, as well as the largest publicly available text-dependent and text-prompted speaker verification database in English, and the largest publicly available evaluation dataset for text-independent speaker verification. The database covers a wide range of ages, genders, and accents. The DeepMine project was launched in early 2017. After completing database design, Android application and server application development, data collection started in mid-2017, was finally completed by the end of 2018, and the cleaned final version was released in early 2019. This database is primarily used for text-dependent speaker verification, and is also suitable for training automatic speech recognition (ASR) models for Persian.