Real Time Speech Emotion Recognition

Recognizes the emotion of the speaker in real time when spoken through a microphone. The emotion is classified in English and Bengali. Made with React, Flask, and Tensorflow.

Usage

Install the dependencies using the following command:

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Then run the following command to start the application:

./start.sh

Or, in windows

.\start.ps1

Visit localhost:8000 to use the application.

Docker

You can also use the docker image to run the application. You need to have docker installed on your system. Pull the image using the following command:

For Windows, install the Docker Desktop from here.

Then start Docker Desktop and run the following commands:

docker pull fo0d/real_time_speech_emotion_recognition:latest

Then run the following command to start the application:

docker run -dp 8000:8000 fo0d/real_time_speech_emotion_recognition

The server needs a few seconds to spin up. After that, visit localhost:8000 through your browser to use the system!

If 8000 is being used by another application, change the port number if you want to use a different port. For example, if you want to use port 5000, then run the following command:

docker run -dp 8000:5000 fo0d/real_time_speech_emotion_recognition

Information

The following pre-trained models were fine-tuned for SER task:

• VGG16
• VGGish
• HuBERT

All the models are re-trained on the following datasets:

• English
  • TESS
  • RAVDESS
  • SAVEE
  • CREMA-D
• Bengali
  • SUBESCO
  • BSER

References

If you use this code, please cite the following paper:

Coming soon...