Khmer ASR

Automatic Speech Recognition

Khmer Automatic Speech Recognition (Khmer ASR) is an application that uses Artificial Intelligence technology to automatically convert speech and audio files in Khmer into written text.


This application was entirely developed by a team of young researchers from Cambodia Academy of Digital Technology (formerly known as National Institute of Posts, Telecommunications, and ICT) of the Ministry of Posts and Telecommunications, using a combination of Statistical Learning and Deep Learning methods, which are the new automation approaches for converting speech into written text for complex script languages ​​with less digital data.

Voice-based Khmer ASR

This is where the users perform realtime speech recognition by using their voice. First, the users need to click the microphone button at the bottom of the screen, then the speech recognition is started. The output transcript will be displayed in the text box and the recording time is shown as in the figure.

After having done speech recognition, the users can copy or delete the output transcript by using button copy and button delete respectively.

Note that the user is required to allow the recording voice permission so that functionality can be taken.

File-based Khmer ASR

This is where the users perform speech recognition via sound file. The application supports four types of sound file such as wav, mp3, m4a, and ogg. To start the function, firstly, the user is required to upload the sound file by clicking the button upload at the right end. Finally, the user can start speech recognition by clicking the button convert in the bottom of the screen. The output transcript will be displayed in the text box and the recording time is shown as in the figure.

After having done speech recognition, the users can copy or delete the output transcript by using button copy and button delete respectively.

Note that the user is required to allow the file permission so that functionality can be taken.

  1. Application Information

    • Application name: Khmer ASR

    • Khmer ASR is developed by research team of Research and Innovation Center, Cambodia Academy Digital Technology. It is officially under the support from Ministry of Posts, Telecoms and ICT in order to provide an Speech-to-Text mobile application for the general use cases, especially for meeting and education.

    • Language: Khmer language

    • This application does not contain ads.

  2. Security

        • This application does not collect any data from the user.

  3. Changes

    • This Privacy Policy may be updated from time to time for any reason. We will notify you of any changes to our Privacy Policy by posting the new Privacy Policy here. You are advised to consult this Privacy Policy regularly for any changes, as continued use is deemed approval of all changes.

  4. Contact Us

If you have any questions regarding the application, please contact us via email at rnd@niptict.edu.kh.


Screenshot

Splash Screen

Application Menu

Voice-based Khmer ASR

File-based Khmer ASR

About Us

Voice Recording Permission Request

File Permission Request