Deep learning for Quranic reciter recognition and audio content identification
Keywords: Speaker identification · Speaker recognition · Deep learning · Transfer learning · Audio embedding
TL;DR: Quran reciter recognition
Abstract: This paper presents a novel approach for identifying the re-
citer, sura, and verse of a given Quranic passage using pre-trained em-
bedding models and transfer learning. Our approach involves training
a deep learning model on a large Quranic recitation audio recordings
dataset and using the resulting embeddings to compare and classify dif-
ferent reciters. We also present a workflow for identifying the specific
sura and verse of a Quranic passage using a speech-to-text model and
elastic-search for the query. We evaluate our approach using a variety
of metrics and demonstrate its effectiveness in accurately identifying the
reciter, sura, and verse of a given passage. We discuss the potential ap-
plications of this approach in the fields of Quranic studies and Islamic
education and outline directions for future work in this area.
Submission Category: Machine learning algorithms
Submission Number: 27
Loading