In this project we record emotions and voice from your personal computer and project those onto Roboy.
For speech-to-text, text-to-speech and emotion detection Microsoft Azure Cognitive Services are used.
For speech-to-text speech_to_text/recognition_node.py
creates a ros2 service at
/roboy/cognition/speech/recognition
.
This project consists of two parts:
- Servers on your personal computer for emotion detection and microphone recording
- Recognition node and main loop which connects to your servers and performs
pyroboy.listen()
,pyroboy.say("bla")
,pyroboy.show_emotion()
.
- Run
pip install -r requirements && pip install -e .
to install the necessary python packages. - Create a file called
secrets.yaml
inside of the configs folder with the following keys and their corresponding values:stt_key, tts_key, face_key, service_region
where the keys are from Microsoft Cognitive Services.
docker-compose up --detach imitator