Here is the 2 mins demo for the application: https://www.youtube.com/watch?v=qg4M_2HnCCA
EmoSense is an innovative tool designed to analyze emotions by leveraging the power of cutting-edge machine learning technologies. It combines the visual emotion recognition capabilities of deepfaceML, the audio transcription accuracy of a speech-to-text ML algorithm, and the advanced understanding and processing abilities of OpenAI's GPT-3 through a command-line wrapper. This unique blend allows EmoSense to provide comprehensive emotion analysis from video recordings, making it an ideal tool for a wide range of applications including mental health assessment, user experience research, and interactive applications.
Visual Emotion Recognition: Utilizes deepfaceML to detect and analyze facial expressions in videos for emotion recognition. Speech-to-Text Transcription: Converts spoken words in videos into text using a state-of-the-art speech-to-text ML algorithm, enhancing the emotion analysis process. Emotion Analysis through GPT-3: Leverages a command-line wrapper for OpenAI's GPT-3 to interpret both visual and textual cues for a holistic understanding of the user's emotional state. Mobile Compatibility: Designed for use in mobile applications, allowing users to record themselves directly within the app for real-time emotion analysis. User-Friendly Interface: Easy-to-use interface for recording videos, with immediate feedback on emotion analysis results.
Python 3.8 or higher Node.js (for the command-line wrapper) OpenAI API key (for GPT-3 integration) Access to a speech-to-text ML API
- Fork and clone the repository
- Navigate to the project directory: cd emosense
- Install the required Python dependencies: pip install -r requirements.txt
- Set up the command-line wrapper for OpenAI's GPT-3 (follow the instructions provided in the cli-wrapper directory).
OpenAI API Key: Store your OpenAI API key in a .env file as follows: OPENAI_API_KEY='your_openai_api_key_here' Speech-to-Text API Key: Similarly, store your speech-to-text API key in the .env file.
streamlit run main.py
Use EmoSense to record yourself speaking about any topic. The app will analyze your facial expressions and speech to provide a comprehensive emotion analysis. This can be particularly useful for mental health tracking, user experience studies, or any application where emotional feedback is valuable.