tags: | |
- speech-recognition | |
language: | |
- ind | |
# indspeech_digit_cdsr | |
INDspeech_DIGIT_CDSR is the first Indonesian speech dataset for connected digit speech recognition (CDSR). The data was developed by TELKOMRisTI (R&D Division, PT Telekomunikasi Indonesia) in collaboration with Advanced Telecommunication Research Institute International (ATR) Japan and Bandung Institute of Technology (ITB) under the Asia-Pacific Telecommunity (APT) project in 2004 [Sakti et al., 2004]. Although it was originally developed for a telecommunication system for hearing and speaking impaired people, it can be used for other applications, i.e., automatic call centers that recognize telephone numbers. | |
## Dataset Usage | |
Run `pip install nusacrowd` before loading the dataset through HuggingFace's `load_dataset`. | |
## Citation | |
``` | |
@inproceedings{sakti-icslp-2004, | |
title = "Indonesian Speech Recognition for Hearing and Speaking Impaired People", | |
author = "Sakti, Sakriani and Hutagaol, Paulus and Arman, Arry Akhmad and Nakamura, Satoshi", | |
booktitle = "Proc. International Conference on Spoken Language Processing (INTERSPEECH - ICSLP)", | |
year = "2004", | |
pages = "1037--1040" | |
address = "Jeju Island, Korea" | |
} | |
``` | |
## License | |
CC-BY-NC-SA-4.0 | |
## Homepage | |
[https://github.com/s-sakti/data_indsp_digit_cdsr](https://github.com/s-sakti/data_indsp_digit_cdsr) | |
### NusaCatalogue | |
For easy indexing and metadata: [https://indonlp.github.io/nusa-catalogue](https://indonlp.github.io/nusa-catalogue) |