Development of a Devanagari Speech Dataset for Audio CAPTCHA Using Mobile-Based Recording and Automated Segmentation
International Journal of Sciences and Innovation Engineering · 05.08.2026 · 3 (8) · 27-38
DOI: 10.70849/ijsci03082690485 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Sciences and Innovation Engineering, online erschienen am 05.08.2026. Verfasst von Dr. Pate Sanjay E..
Abstract (Originalsprache)
The development of a scalable and well-organized Devanagari speech database for audio CAPTCHA application is presented in this research. 536 recordings of 10 digits and 23 chosen Devanagari characters were gathered from a variety of speakers, including boys and girls, using mobile devices to produce the dataset. For audio format conversion (M4A to MP3), pause-based segmentation, and methodical arrangement of audio samples, a Python-based processing pipeline is put into place. A total of 10,675 verified audio samples are produced, comprising 3,567-digit and 7,108-character samples. By selecting properly spoken samples and eliminating incorrect or ambiguous recordings, manual verification is carried out to guarantee data quality. The proposed approach offers an affordable, scalable, and useful way to create speech datasets in language environments with limited resources. Applications like assistive technology, voice recognition, and audio CAPTCHA systems can use the created dataset. The novel dataset presented in this work is accessible to the public on: 1. Mendeley Data at https://data.mendeley.com/datasets/cwd42xmrnp/1 2. The IEEE data port is available at https://ieee-dataport.org/documents/natural-devanagari-character-digit-audio-dataset-deva-audio.
Zitieren
APA
E., D. P. S. (2026). Development of a Devanagari Speech Dataset for Audio CAPTCHA Using Mobile-Based Recording and Automated Segmentation. International Journal of Sciences and Innovation Engineering, 3(8), 27-38. https://doi.org/10.70849/ijsci03082690485
Vancouver
E. DPS. Development of a Devanagari Speech Dataset for Audio CAPTCHA Using Mobile-Based Recording and Automated Segmentation. International Journal of Sciences and Innovation Engineering. 2026;3(8):27-38. doi:10.70849/ijsci03082690485
BibTeX
@article{E2026,
author = {E., Dr. Pate Sanjay},
title = {Development of a Devanagari Speech Dataset for Audio CAPTCHA Using Mobile-Based Recording and Automated Segmentation},
journal = {International Journal of Sciences and Innovation Engineering},
year = {2026},
volume = {3},
number = {8},
pages = {27--38},
doi = {10.70849/ijsci03082690485}
}
EUVIMED