arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2305.07034eess.AScs.AIcs.CLcs.LGcs.SD
收藏

使用端到端深度学习进行《古兰经》诵读识别

Quran Recitation Recognition using End-to-End Deep Learning

Arab International University(阿拉伯国际大学)

Ahmad Al Harere, Khloud Al Jallad

AI总结:

本文提出一种基于CNN-双向GRU和CTC的端到端深度学习模型,用于《古兰经》诵读识别,在公共数据集Ar-DAD上取得8.34%的词错误率和2.42%的字符错误率,为后续研究提供基线。

摘要:

《古兰经》是伊斯兰教的圣典,其诵读是宗教中的一个重要方面。自动识别《古兰经》的诵读是一项具有挑战性的任务,因为其独特的规则不适用于正常说话的语音。在该领域已进行了大量研究,但以往的工作要么将诵读错误检测作为分类任务,要么使用传统的自动语音识别(ASR)。在本文中,我们提出了一种新颖的端到端深度学习模型来识别《古兰经》的诵读。所提出的模型是一个CNN-双向GRU编码器,使用CTC作为目标函数,以及一个基于字符的波束搜索解码器。此外,以往所有工作都是在包含短经文和少数几章的小型私有数据集上进行的。由于使用私有数据集,没有进行过比较。为解决此问题,我们使用了最近发布的一个公共数据集(Ar-DAD),该数据集包含约37章,由30位诵读者以不同的诵读速度和不同类型的发音规则进行诵读。所提出模型的性能使用语音识别中最常见的评估指标进行评估:词错误率(WER)和字符错误率(CER)。结果为WER 8.34%和CER 2.42%。我们希望这项研究能成为未来在该公共新数据集(Ar-DAD)上进行研究比较的基线。

英文摘要:

The Quran is the holy scripture of Islam, and its recitation is an important aspect of the religion. Recognizing the recitation of the Holy Quran automatically is a challenging task due to its unique rules that are not applied in normal speaking speeches. A lot of research has been done in this domain, but previous works have detected recitation errors as a classification task or used traditional automatic speech recognition (ASR). In this paper, we proposed a novel end-to-end deep learning model for recognizing the recitation of the Holy Quran. The proposed model is a CNN-Bidirectional GRU encoder that uses CTC as an objective function, and a character-based decoder which is a beam search decoder. Moreover, all previous works were done on small private datasets consisting of short verses and a few chapters of the Holy Quran. As a result of using private datasets, no comparisons were done. To overcome this issue, we used a public dataset that has recently been published (Ar-DAD) and contains about 37 chapters that were recited by 30 reciters, with different recitation speeds and different types of pronunciation rules. The proposed model performance was evaluated using the most common evaluation metrics in speech recognition, word error rate (WER), and character error rate (CER). The results were 8.34% WER and 2.42% CER. We hope this research will be a baseline for comparisons with future research on this public new dataset (Ar-DAD).

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑