| 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) |