Bengali Real Number Speech Corpus
收藏arXiv2018-03-27 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/1803.10136v1
下载链接
链接失效反馈官方服务：
资源简介：
本研究开发了首个全面的孟加拉语实数语音数据集，名为‘Bengali Real Number Speech Corpus’。该数据集由来自不同地区的10名孟加拉语母语者录制，包含超过2302个语音样本，总时长近4小时。数据集的创建过程包括随机生成包含孟加拉语实数的字符串，并通过多个录音环境进行录制和过滤。该数据集主要用于自动语音识别（ASR）领域，旨在解决孟加拉语语音识别中数据不足的问题。
This study develops the first comprehensive Bengali real number speech dataset, named "Bengali Real Number Speech Corpus". This dataset was recorded by 10 native Bengali speakers from diverse regions, containing over 2302 speech samples with a total duration of nearly 4 hours. The construction process of this dataset includes randomly generating character strings containing Bengali real numbers, followed by recording and filtering across multiple recording environments. This dataset is primarily used in the field of automatic speech recognition (ASR), aiming to address the issue of insufficient data for Bengali speech recognition.
提供机构：
计算机科学与工程系，沙贾拉尔科技大学
创建时间：
2018-03-27
