WSCoach

A wearable assistant that uses auditory feedback to help users avoid unwanted words in everyday conversation.

Publications

• WSCoach: Wearable Real-time Auditory Feedback for Reducing Unwanted Words in Daily Communication, IMWUT'25
  • Camera Ready PDF

@article{youpeng_wscoach_2025,
	title = { {WSCoach}: {Wearable} {Real-time} {Auditory} {Feedback} for {Reducing} {Unwanted} {Words} in {Daily} {Communication}},
	shorttitle = {{WSCoach}},
	journal = {Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies},
	author = {Youpeng, Zhang and Janaka, Nuwan and Ram, Ashwin and Yin, Peilin and Yang, Tian and Zhao, Shengdong and Dragicevic, Pierre},
	year = {2025},
    doi = {10.1145/3749531},
}

Contact person

• Zhang Youpeng
• Nuwan Janaka (In)

Project links

• Project folder (private access): here

Table of Contents

• WSCOACH
  • Table of Contents
    • Features
    • Installation
    • Usage
  • FasterWhisperModel Module
    • Features
    • Installation
    • Usage
  • Record Module
    • Features
    • Installation
    • Usage
  • TTS Module
    • Features
    • Installation
    • Usage
  • Contributing
  • License

Features

• Word Detection: Recognizes specific words or phrases in a given text string.
• Tracking: Keeps track of the number of times each word is detected.

Installation

• Make sure Python3 (>=3.10) is installed.
• Install the required dependencies in requirement.txt (e.g., pip install -r requirement.txt).

Usage

• Run the HTML_render.py (python HTML_render.py) to use the basic features of WSCoach.

FasterWhisperModel Module

The FasterWhisperModel module provides a wrapper for the Faster-Whisper ASR (Automatic Speech Recognition) model. It allows for efficient transcription of audio files or arrays, with a focus on faster processing.

Features

• Fast Transcription: Utilizes the Faster-Whisper ASR model for efficient and quick audio transcription.
• File and Array Transcription: Supports transcription from both audio files and arrays.

Installation

No additional installation is required for this module. However, make sure that the required dependencies are installed in your Python environment.

Usage

1. Import the FasterWhisperModel class into your Python script.

   from faster_whisper import FasterWhisperModel

2. Create an instance of the FasterWhisperModel class.

   model = FasterWhisperModel()

3. Use the TranscribeFile method to transcribe audio from a file.

   text = model.TranscribeFile("./path/to/audio/file.wav")
   print(f"Transcribed Text: {text}")

4. Use the TranscribeArray method to transcribe audio from a NumPy array.

   text = model.TranscribeArray(audio_array)
   print(f"Transcribed Text: {text}")

Record Module

The Record module provides functionality for recording audio using the PyAudio library. It captures audio input from a specified input device and writes it to a WAV file. Additionally, it supports playing back recorded audio.

Features

• Audio Recording: Records audio input from a specified input device.
• Buffer Management: Manages an internal buffer to store audio data efficiently.
• Wave File Creation: Writes recorded audio to a WAV file.
• Audio Playback: Supports playback of recorded audio.

Installation

No additional installation is required for this module. However, make sure that the required dependencies are installed in your Python environment.

Usage

1. Import the Record class into your Python script.

   from Record import Record

2. Create an instance of the Record class.

   record = Record()

3. Optionally, set the input and output devices using the provided methods.

   input_device_names = record.get_device_input_names()
   output_device_names = record.get_device_output_names()
   
   record.set_device_input_index("Your_Input_Device_Name")
   record.set_device_output_index("Your_Output_Device_Name")

4. Start the recording thread.

   record.start()

5. Optionally, capture the audio buffer.

   audio_buffer = record.GetBuffer()

6. Optionally, end the recording.

   record.endRecord()

7. Optionally, play the recorded audio.

   record.playWave("Your_Wave_File_Name", detect=detect_instance, first_name="Your_First_Name")

TTS Module

The TTS module utilizes the pyttsx4 library to convert text into speech and save the generated audio as WAV files. It provides the ability to adjust the speech rate and save individual words or phrases as separate audio files.

Features

• Text-to-Speech Conversion: Converts input text into speech.
• Adjustable Speech Rate: Allows the user to set the speech rate.
• Saving to WAV Files: Saves the generated speech as WAV files.

Installation

No additional installation is required for this module. However, make sure that the required dependencies are installed in your Python environment.

Usage

1. Import the TTS class into your Python script.

   from TTS import TTS

2. Create an instance of the TTS class.

   tts = TTS()

3. Optionally, set the speech rate.

   tts.SetRate(2.0)

4. Use the Word2File method to convert and save text as a WAV file.

   tts.Word2File("Hello World")

Contributing

If you would like to contribute to the development of this project or report issues, please contract us.

License

This project is open source and available under the MIT License.